diff --git a/liquid_clustering_analysis_20250815_092450.json b/liquid_clustering_analysis_20250815_092450.json
deleted file mode 100644
index 06fad06..0000000
--- a/liquid_clustering_analysis_20250815_092450.json
+++ /dev/null
@@ -1,202 +0,0 @@
-{
-  "llm_analysis": "❌ Failed to obtain Databricks token. Please set the environment variable DATABRICKS_TOKEN.",
-  "extracted_data": {
-    "filter_columns": [
-      {
-        "expression": "(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_sold_date_sk IS NOT NULL)",
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC"
-      },
-      {
-        "expression": "(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date IS NOT NULL)",
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.date_dim",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC"
-      },
-      {
-        "expression": "(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date >= DATE '1990-01-02')",
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.date_dim",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC"
-      },
-      {
-        "expression": "(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk IS NOT NULL)",
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.date_dim",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC"
-      }
-    ],
-    "join_columns": [
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_sold_date_sk",
-        "key_type": "LEFT_KEYS",
-        "node_name": "Left Semi Join",
-        "node_tag": "PHOTON_BROADCAST_HASH_JOIN_EXEC"
-      },
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk",
-        "key_type": "RIGHT_KEYS",
-        "node_name": "Left Semi Join",
-        "node_tag": "PHOTON_BROADCAST_HASH_JOIN_EXEC"
-      }
-    ],
-    "groupby_columns": [
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      }
-    ],
-    "aggregate_columns": [
-      {
-        "expression": "avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price))",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit))",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price))",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit))",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      }
-    ],
-    "table_info": {
-      "tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo": {
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC",
-        "node_id": "6224",
-        "current_clustering_keys": [
-          "cs_sold_date_sk"
-        ],
-        "table_size_gb": 1279.1372904106975,
-        "files_read_bytes": 1373463207352,
-        "files_pruned_bytes": 4635328390,
-        "io_read_bytes": 170792875003,
-        "total_scan_gb": 1283.4542763810605,
-        "size_classification": "large"
-      },
-      "tpcds.tpcds_sf10000_delta_lc.date_dim": {
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.date_dim",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC",
-        "node_id": "6016",
-        "current_clustering_keys": [
-          "d_date_sk"
-        ],
-        "table_size_gb": 0.0017275810241699219,
-        "files_read_bytes": 1854976,
-        "files_pruned_bytes": 0,
-        "io_read_bytes": 846678,
-        "total_scan_gb": 0.0017275810241699219,
-        "size_classification": "small"
-      }
-    },
-    "scan_nodes": [],
-    "join_nodes": [],
-    "filter_nodes": [],
-    "metadata_summary": {
-      "total_nodes": 30,
-      "total_graphs": 1,
-      "filter_expressions_count": 4,
-      "join_expressions_count": 2,
-      "groupby_expressions_count": 3,
-      "aggregate_expressions_count": 16,
-      "tables_identified": 2,
-      "scan_nodes_count": 0,
-      "join_nodes_count": 0,
-      "filter_nodes_count": 0
-    }
-  },
-  "performance_context": {
-    "total_time_sec": 253.607,
-    "read_gb": 0,
-    "rows_produced": 10000,
-    "rows_read": 14327959412,
-    "data_selectivity": 0.051481078263833586
-  },
-  "summary": {
-    "analysis_method": "LLM-based",
-    "tables_identified": 2,
-    "total_filter_columns": 4,
-    "total_join_columns": 2,
-    "total_groupby_columns": 3,
-    "total_aggregate_columns": 16,
-    "scan_nodes_count": 0,
-    "llm_provider": "databricks"
-  }
-}
\ No newline at end of file
diff --git a/liquid_clustering_analysis_20250815_092450.md b/liquid_clustering_analysis_20250815_092450.md
deleted file mode 100644
index f711e07..0000000
--- a/liquid_clustering_analysis_20250815_092450.md
+++ /dev/null
@@ -1,85 +0,0 @@
-# Liquid Clustering Analysis Report
-
-**Generated Date**: 2025-08-15 09:24:50  
-**Analysis Method**: LLM-based  
-**LLM Provider**: databricks
-
-## 📊 Performance Overview
-
-| Item | Value |
-|------|-----|
-| Execution Time | 253.6 seconds |
-| Data Read | 0.00GB |
-| Output Rows | 10,000 rows |
-| Read Rows | 14,327,959,412 rows |
-| Filter Rate | 0.0515 |
-
-## 🔍 Extracted Metadata
-
-### Filter Conditions (4 items)
-1. `(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_sold_date_sk IS NOT NULL)` (ノード: Scan tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo)
-2. `(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date IS NOT NULL)` (ノード: Scan tpcds.tpcds_sf10000_delta_lc.date_dim)
-3. `(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date >= DATE '1990-01-02')` (ノード: Scan tpcds.tpcds_sf10000_delta_lc.date_dim)
-4. `(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk IS NOT NULL)` (ノード: Scan tpcds.tpcds_sf10000_delta_lc.date_dim)
-
-### JOIN条件 (2個)
-1. `tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_sold_date_sk` (LEFT_KEYS)
-2. `tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk` (RIGHT_KEYS)
-
-### GROUP BY条件 (3個)
-1. `tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk` (ノード: Grouping Aggregate)
-2. `tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk` (ノード: Grouping Aggregate)
-3. `tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk` (ノード: Grouping Aggregate)
-
-### 集約関数 (16個)
-1. `avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price))` (ノード: Grouping Aggregate)
-2. `min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)` (ノード: Grouping Aggregate)
-3. `max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)` (ノード: Grouping Aggregate)
-4. `count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)` (ノード: Grouping Aggregate)
-5. `avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit))` (ノード: Grouping Aggregate)
-6. `min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)` (ノード: Grouping Aggregate)
-7. `max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)` (ノード: Grouping Aggregate)
-8. `count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)` (ノード: Grouping Aggregate)
-9. `avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price))` (ノード: Grouping Aggregate)
-10. `min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)` (ノード: Grouping Aggregate)
-... 他 6個
-
-## 🏷️ 識別されたテーブル (2個)
-
-- **tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo** (ノード: Scan tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo)
-  - 現在のクラスタリングキー: `cs_sold_date_sk`
-- **tpcds.tpcds_sf10000_delta_lc.date_dim** (ノード: Scan tpcds.tpcds_sf10000_delta_lc.date_dim)
-  - 現在のクラスタリングキー: `d_date_sk`
-
-## 🔎 スキャンノード分析 (0個)
-
-
-## 🤖 LLM分析結果
-
-❌ Failed to obtain Databricks token. Please set the environment variable DATABRICKS_TOKEN.
-
-#### キー選定の原則
-- **基本原則**: フィルタ列での読み取り最適化（データスキッピング）を優先
-- **優先順位**: 「よく絞り込みに使う列」を第一優先に選定
-
-#### GROUP BY キーの考慮条件
-
-1. フィルタにも使う列がGROUP BYにも登場する場合
-2. シャッフルに乗る中間データ量の削減が見込める場合
-3. キーのカーディナリティが低〜中程度で極端なスキューが少ない場合
-
-#### 実務上の推奨
-
-上記条件を満たさない場合は、常にフィルタ列を優先
-
-## 📋 分析サマリー
-
-- **分析対象テーブル数**: 2
-- **フィルター条件数**: 4
-- **JOIN条件数**: 2
-- **GROUP BY条件数**: 3
-- **集約関数数**: 16
-- **スキャンノード数**: 0
-
----
-*Report generation time: 2025-08-15 09:24:50*
diff --git a/liquid_clustering_analysis_20250815_092456.json b/liquid_clustering_analysis_20250815_092456.json
deleted file mode 100644
index 06fad06..0000000
--- a/liquid_clustering_analysis_20250815_092456.json
+++ /dev/null
@@ -1,202 +0,0 @@
-{
-  "llm_analysis": "❌ Failed to obtain Databricks token. Please set the environment variable DATABRICKS_TOKEN.",
-  "extracted_data": {
-    "filter_columns": [
-      {
-        "expression": "(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_sold_date_sk IS NOT NULL)",
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC"
-      },
-      {
-        "expression": "(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date IS NOT NULL)",
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.date_dim",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC"
-      },
-      {
-        "expression": "(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date >= DATE '1990-01-02')",
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.date_dim",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC"
-      },
-      {
-        "expression": "(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk IS NOT NULL)",
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.date_dim",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC"
-      }
-    ],
-    "join_columns": [
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_sold_date_sk",
-        "key_type": "LEFT_KEYS",
-        "node_name": "Left Semi Join",
-        "node_tag": "PHOTON_BROADCAST_HASH_JOIN_EXEC"
-      },
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk",
-        "key_type": "RIGHT_KEYS",
-        "node_name": "Left Semi Join",
-        "node_tag": "PHOTON_BROADCAST_HASH_JOIN_EXEC"
-      }
-    ],
-    "groupby_columns": [
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      }
-    ],
-    "aggregate_columns": [
-      {
-        "expression": "avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price))",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit))",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price))",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit))",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      }
-    ],
-    "table_info": {
-      "tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo": {
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC",
-        "node_id": "6224",
-        "current_clustering_keys": [
-          "cs_sold_date_sk"
-        ],
-        "table_size_gb": 1279.1372904106975,
-        "files_read_bytes": 1373463207352,
-        "files_pruned_bytes": 4635328390,
-        "io_read_bytes": 170792875003,
-        "total_scan_gb": 1283.4542763810605,
-        "size_classification": "large"
-      },
-      "tpcds.tpcds_sf10000_delta_lc.date_dim": {
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.date_dim",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC",
-        "node_id": "6016",
-        "current_clustering_keys": [
-          "d_date_sk"
-        ],
-        "table_size_gb": 0.0017275810241699219,
-        "files_read_bytes": 1854976,
-        "files_pruned_bytes": 0,
-        "io_read_bytes": 846678,
-        "total_scan_gb": 0.0017275810241699219,
-        "size_classification": "small"
-      }
-    },
-    "scan_nodes": [],
-    "join_nodes": [],
-    "filter_nodes": [],
-    "metadata_summary": {
-      "total_nodes": 30,
-      "total_graphs": 1,
-      "filter_expressions_count": 4,
-      "join_expressions_count": 2,
-      "groupby_expressions_count": 3,
-      "aggregate_expressions_count": 16,
-      "tables_identified": 2,
-      "scan_nodes_count": 0,
-      "join_nodes_count": 0,
-      "filter_nodes_count": 0
-    }
-  },
-  "performance_context": {
-    "total_time_sec": 253.607,
-    "read_gb": 0,
-    "rows_produced": 10000,
-    "rows_read": 14327959412,
-    "data_selectivity": 0.051481078263833586
-  },
-  "summary": {
-    "analysis_method": "LLM-based",
-    "tables_identified": 2,
-    "total_filter_columns": 4,
-    "total_join_columns": 2,
-    "total_groupby_columns": 3,
-    "total_aggregate_columns": 16,
-    "scan_nodes_count": 0,
-    "llm_provider": "databricks"
-  }
-}
\ No newline at end of file
diff --git a/liquid_clustering_analysis_20250815_092456.md b/liquid_clustering_analysis_20250815_092456.md
deleted file mode 100644
index 874a794..0000000
--- a/liquid_clustering_analysis_20250815_092456.md
+++ /dev/null
@@ -1,85 +0,0 @@
-# Liquid Clustering Analysis Report
-
-**Generated Date**: 2025-08-15 09:24:56  
-**Analysis Method**: LLM-based  
-**LLM Provider**: databricks
-
-## 📊 Performance Overview
-
-| Item | Value |
-|------|-----|
-| Execution Time | 253.6 seconds |
-| Data Read | 0.00GB |
-| Output Rows | 10,000 rows |
-| Read Rows | 14,327,959,412 rows |
-| Filter Rate | 0.0515 |
-
-## 🔍 Extracted Metadata
-
-### Filter Conditions (4 items)
-1. `(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_sold_date_sk IS NOT NULL)` (ノード: Scan tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo)
-2. `(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date IS NOT NULL)` (ノード: Scan tpcds.tpcds_sf10000_delta_lc.date_dim)
-3. `(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date >= DATE '1990-01-02')` (ノード: Scan tpcds.tpcds_sf10000_delta_lc.date_dim)
-4. `(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk IS NOT NULL)` (ノード: Scan tpcds.tpcds_sf10000_delta_lc.date_dim)
-
-### JOIN条件 (2個)
-1. `tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_sold_date_sk` (LEFT_KEYS)
-2. `tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk` (RIGHT_KEYS)
-
-### GROUP BY条件 (3個)
-1. `tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk` (ノード: Grouping Aggregate)
-2. `tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk` (ノード: Grouping Aggregate)
-3. `tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk` (ノード: Grouping Aggregate)
-
-### 集約関数 (16個)
-1. `avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price))` (ノード: Grouping Aggregate)
-2. `min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)` (ノード: Grouping Aggregate)
-3. `max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)` (ノード: Grouping Aggregate)
-4. `count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)` (ノード: Grouping Aggregate)
-5. `avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit))` (ノード: Grouping Aggregate)
-6. `min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)` (ノード: Grouping Aggregate)
-7. `max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)` (ノード: Grouping Aggregate)
-8. `count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)` (ノード: Grouping Aggregate)
-9. `avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price))` (ノード: Grouping Aggregate)
-10. `min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)` (ノード: Grouping Aggregate)
-... 他 6個
-
-## 🏷️ 識別されたテーブル (2個)
-
-- **tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo** (ノード: Scan tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo)
-  - 現在のクラスタリングキー: `cs_sold_date_sk`
-- **tpcds.tpcds_sf10000_delta_lc.date_dim** (ノード: Scan tpcds.tpcds_sf10000_delta_lc.date_dim)
-  - 現在のクラスタリングキー: `d_date_sk`
-
-## 🔎 スキャンノード分析 (0個)
-
-
-## 🤖 LLM分析結果
-
-❌ Failed to obtain Databricks token. Please set the environment variable DATABRICKS_TOKEN.
-
-#### キー選定の原則
-- **基本原則**: フィルタ列での読み取り最適化（データスキッピング）を優先
-- **優先順位**: 「よく絞り込みに使う列」を第一優先に選定
-
-#### GROUP BY キーの考慮条件
-
-1. フィルタにも使う列がGROUP BYにも登場する場合
-2. シャッフルに乗る中間データ量の削減が見込める場合
-3. キーのカーディナリティが低〜中程度で極端なスキューが少ない場合
-
-#### 実務上の推奨
-
-上記条件を満たさない場合は、常にフィルタ列を優先
-
-## 📋 分析サマリー
-
-- **分析対象テーブル数**: 2
-- **フィルター条件数**: 4
-- **JOIN条件数**: 2
-- **GROUP BY条件数**: 3
-- **集約関数数**: 16
-- **スキャンノード数**: 0
-
----
-*Report generation time: 2025-08-15 09:24:56*
diff --git a/liquid_clustering_analysis_20250815_092500.json b/liquid_clustering_analysis_20250815_092500.json
deleted file mode 100644
index 06fad06..0000000
--- a/liquid_clustering_analysis_20250815_092500.json
+++ /dev/null
@@ -1,202 +0,0 @@
-{
-  "llm_analysis": "❌ Failed to obtain Databricks token. Please set the environment variable DATABRICKS_TOKEN.",
-  "extracted_data": {
-    "filter_columns": [
-      {
-        "expression": "(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_sold_date_sk IS NOT NULL)",
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC"
-      },
-      {
-        "expression": "(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date IS NOT NULL)",
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.date_dim",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC"
-      },
-      {
-        "expression": "(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date >= DATE '1990-01-02')",
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.date_dim",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC"
-      },
-      {
-        "expression": "(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk IS NOT NULL)",
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.date_dim",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC"
-      }
-    ],
-    "join_columns": [
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_sold_date_sk",
-        "key_type": "LEFT_KEYS",
-        "node_name": "Left Semi Join",
-        "node_tag": "PHOTON_BROADCAST_HASH_JOIN_EXEC"
-      },
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk",
-        "key_type": "RIGHT_KEYS",
-        "node_name": "Left Semi Join",
-        "node_tag": "PHOTON_BROADCAST_HASH_JOIN_EXEC"
-      }
-    ],
-    "groupby_columns": [
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      }
-    ],
-    "aggregate_columns": [
-      {
-        "expression": "avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price))",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit))",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price))",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit))",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      },
-      {
-        "expression": "count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)",
-        "node_name": "Grouping Aggregate",
-        "node_tag": "PHOTON_GROUPING_AGG_EXEC"
-      }
-    ],
-    "table_info": {
-      "tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo": {
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC",
-        "node_id": "6224",
-        "current_clustering_keys": [
-          "cs_sold_date_sk"
-        ],
-        "table_size_gb": 1279.1372904106975,
-        "files_read_bytes": 1373463207352,
-        "files_pruned_bytes": 4635328390,
-        "io_read_bytes": 170792875003,
-        "total_scan_gb": 1283.4542763810605,
-        "size_classification": "large"
-      },
-      "tpcds.tpcds_sf10000_delta_lc.date_dim": {
-        "node_name": "Scan tpcds.tpcds_sf10000_delta_lc.date_dim",
-        "node_tag": "UNKNOWN_DATA_SOURCE_SCAN_EXEC",
-        "node_id": "6016",
-        "current_clustering_keys": [
-          "d_date_sk"
-        ],
-        "table_size_gb": 0.0017275810241699219,
-        "files_read_bytes": 1854976,
-        "files_pruned_bytes": 0,
-        "io_read_bytes": 846678,
-        "total_scan_gb": 0.0017275810241699219,
-        "size_classification": "small"
-      }
-    },
-    "scan_nodes": [],
-    "join_nodes": [],
-    "filter_nodes": [],
-    "metadata_summary": {
-      "total_nodes": 30,
-      "total_graphs": 1,
-      "filter_expressions_count": 4,
-      "join_expressions_count": 2,
-      "groupby_expressions_count": 3,
-      "aggregate_expressions_count": 16,
-      "tables_identified": 2,
-      "scan_nodes_count": 0,
-      "join_nodes_count": 0,
-      "filter_nodes_count": 0
-    }
-  },
-  "performance_context": {
-    "total_time_sec": 253.607,
-    "read_gb": 0,
-    "rows_produced": 10000,
-    "rows_read": 14327959412,
-    "data_selectivity": 0.051481078263833586
-  },
-  "summary": {
-    "analysis_method": "LLM-based",
-    "tables_identified": 2,
-    "total_filter_columns": 4,
-    "total_join_columns": 2,
-    "total_groupby_columns": 3,
-    "total_aggregate_columns": 16,
-    "scan_nodes_count": 0,
-    "llm_provider": "databricks"
-  }
-}
\ No newline at end of file
diff --git a/liquid_clustering_analysis_20250815_092500.md b/liquid_clustering_analysis_20250815_092500.md
deleted file mode 100644
index e6d230a..0000000
--- a/liquid_clustering_analysis_20250815_092500.md
+++ /dev/null
@@ -1,85 +0,0 @@
-# Liquid Clustering Analysis Report
-
-**Generated Date**: 2025-08-15 09:25:00  
-**Analysis Method**: LLM-based  
-**LLM Provider**: databricks
-
-## 📊 Performance Overview
-
-| Item | Value |
-|------|-----|
-| Execution Time | 253.6 seconds |
-| Data Read | 0.00GB |
-| Output Rows | 10,000 rows |
-| Read Rows | 14,327,959,412 rows |
-| Filter Rate | 0.0515 |
-
-## 🔍 Extracted Metadata
-
-### Filter Conditions (4 items)
-1. `(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_sold_date_sk IS NOT NULL)` (ノード: Scan tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo)
-2. `(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date IS NOT NULL)` (ノード: Scan tpcds.tpcds_sf10000_delta_lc.date_dim)
-3. `(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date >= DATE '1990-01-02')` (ノード: Scan tpcds.tpcds_sf10000_delta_lc.date_dim)
-4. `(tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk IS NOT NULL)` (ノード: Scan tpcds.tpcds_sf10000_delta_lc.date_dim)
-
-### JOIN条件 (2個)
-1. `tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_sold_date_sk` (LEFT_KEYS)
-2. `tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk` (RIGHT_KEYS)
-
-### GROUP BY条件 (3個)
-1. `tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk` (ノード: Grouping Aggregate)
-2. `tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk` (ノード: Grouping Aggregate)
-3. `tpcds.tpcds_sf10000_delta_lc.date_dim.d_date_sk` (ノード: Grouping Aggregate)
-
-### 集約関数 (16個)
-1. `avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price))` (ノード: Grouping Aggregate)
-2. `min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)` (ノード: Grouping Aggregate)
-3. `max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)` (ノード: Grouping Aggregate)
-4. `count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)` (ノード: Grouping Aggregate)
-5. `avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit))` (ノード: Grouping Aggregate)
-6. `min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)` (ノード: Grouping Aggregate)
-7. `max(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)` (ノード: Grouping Aggregate)
-8. `count(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_net_profit)` (ノード: Grouping Aggregate)
-9. `avg(unscaledvalue(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price))` (ノード: Grouping Aggregate)
-10. `min(tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_ext_sales_price)` (ノード: Grouping Aggregate)
-... 他 6個
-
-## 🏷️ 識別されたテーブル (2個)
-
-- **tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo** (ノード: Scan tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo)
-  - 現在のクラスタリングキー: `cs_sold_date_sk`
-- **tpcds.tpcds_sf10000_delta_lc.date_dim** (ノード: Scan tpcds.tpcds_sf10000_delta_lc.date_dim)
-  - 現在のクラスタリングキー: `d_date_sk`
-
-## 🔎 スキャンノード分析 (0個)
-
-
-## 🤖 LLM分析結果
-
-❌ Failed to obtain Databricks token. Please set the environment variable DATABRICKS_TOKEN.
-
-#### キー選定の原則
-- **基本原則**: フィルタ列での読み取り最適化（データスキッピング）を優先
-- **優先順位**: 「よく絞り込みに使う列」を第一優先に選定
-
-#### GROUP BY キーの考慮条件
-
-1. フィルタにも使う列がGROUP BYにも登場する場合
-2. シャッフルに乗る中間データ量の削減が見込める場合
-3. キーのカーディナリティが低〜中程度で極端なスキューが少ない場合
-
-#### 実務上の推奨
-
-上記条件を満たさない場合は、常にフィルタ列を優先
-
-## 📋 分析サマリー
-
-- **分析対象テーブル数**: 2
-- **フィルター条件数**: 4
-- **JOIN条件数**: 2
-- **GROUP BY条件数**: 3
-- **集約関数数**: 16
-- **スキャンノード数**: 0
-
----
-*Report generation time: 2025-08-15 09:25:00*
diff --git a/liquid_clustering_implementation_20250815_092450.sql b/liquid_clustering_implementation_20250815_092450.sql
deleted file mode 100644
index aae8a19..0000000
--- a/liquid_clustering_implementation_20250815_092450.sql
+++ /dev/null
@@ -1,120 +0,0 @@
--- =====================================================
--- Liquid Clustering 実装SQL例
--- 生成日時: 2025-08-15 09:24:50
--- =====================================================
-
--- 【重要】
--- 以下のSQL例は分析結果に基づく推奨事項です。
--- 実際の実装前に、テーブル構造やデータ特性を確認してください。
-
-
--- =====================================================
--- テーブル: tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo
--- 現在のクラスタリングキー: cs_sold_date_sk
--- =====================================================
-
--- 既存テーブルにLiquid Clusteringを適用する場合:
--- ALTER TABLE tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo CLUSTER BY (column1, column2, column3, column4);
-
--- 新規テーブル作成時にLiquid Clusteringを設定する場合:
--- CREATE TABLE tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo_clustered
--- CLUSTER BY (column1, column2, column3, column4)
--- AS SELECT * FROM tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo;
-
--- Delta Live Tablesでの設定例:
--- @dlt.table(
---   cluster_by=["column1", "column2", "column3", "column4"]
--- )
--- def catalog_sales_demo_clustered():
---   return spark.table("tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo")
-
--- クラスタリング状況の確認:
--- DESCRIBE DETAIL tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo;
-
--- クラスタリング統計の確認:
--- ANALYZE TABLE tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo COMPUTE STATISTICS FOR ALL COLUMNS;
-
-
--- =====================================================
--- テーブル: tpcds.tpcds_sf10000_delta_lc.date_dim
--- 現在のクラスタリングキー: d_date_sk
--- =====================================================
-
--- 既存テーブルにLiquid Clusteringを適用する場合:
--- ALTER TABLE tpcds.tpcds_sf10000_delta_lc.date_dim CLUSTER BY (column1, column2, column3, column4);
-
--- 新規テーブル作成時にLiquid Clusteringを設定する場合:
--- CREATE TABLE tpcds.tpcds_sf10000_delta_lc.date_dim_clustered
--- CLUSTER BY (column1, column2, column3, column4)
--- AS SELECT * FROM tpcds.tpcds_sf10000_delta_lc.date_dim;
-
--- Delta Live Tablesでの設定例:
--- @dlt.table(
---   cluster_by=["column1", "column2", "column3", "column4"]
--- )
--- def date_dim_clustered():
---   return spark.table("tpcds.tpcds_sf10000_delta_lc.date_dim")
-
--- クラスタリング状況の確認:
--- DESCRIBE DETAIL tpcds.tpcds_sf10000_delta_lc.date_dim;
-
--- クラスタリング統計の確認:
--- ANALYZE TABLE tpcds.tpcds_sf10000_delta_lc.date_dim COMPUTE STATISTICS FOR ALL COLUMNS;
-
-
--- =====================================================
--- 一般的なLiquid Clustering実装パターン
--- =====================================================
-
--- パターン1: フィルター頻度の高いカラムを優先
--- 推奨順序: 1) フィルター条件カラム 2) JOIN条件カラム 3) GROUP BYカラム
-
--- パターン2: カーディナリティを考慮した順序
--- 低カーディナリティ → 高カーディナリティの順で配置
-
--- パターン3: データアクセスパターンに基づく配置
--- よく一緒に使用されるカラムを近い位置に配置
-
--- =====================================================
--- 実装後のパフォーマンス検証SQL
--- =====================================================
-
--- 1. クエリ実行計画の確認
--- EXPLAIN SELECT ... FROM table_name WHERE ...;
-
--- 2. ファイルスキップ統計の確認
--- SELECT * FROM table_name WHERE filter_column = 'value';
--- -- SQLプロファイラーでファイルスキップ数を確認
-
--- 3. データ配置の確認
--- SELECT 
---   file_path,
---   count(*) as row_count,
---   min(cluster_column1) as min_val,
---   max(cluster_column1) as max_val
--- FROM table_name
--- GROUP BY file_path
--- ORDER BY file_path;
-
--- =====================================================
--- 注意事項
--- =====================================================
-
--- 1. Liquid Clusteringは最大4カラムまで指定可能
--- 2. パーティショニングとは併用不可
--- 3. 既存のZORDER BYは自動的に無効化される
--- 4. クラスタリングの効果は時間とともに向上する（OPTIMIZE実行で最適化）
--- 5. 定期的なOPTIMIZE実行を推奨
--- 6. **重要**: カラムの指定順序はパフォーマンスに影響しません
---    * CLUSTER BY (col1, col2, col3) と CLUSTER BY (col3, col1, col2) は同等
---    * 従来のパーティショニングやZ-ORDERとは異なる重要な特性
-
--- OPTIMIZE実行例:
--- OPTIMIZE table_name;
-
--- =====================================================
--- 生成情報
--- =====================================================
--- 生成日時: 2025-08-15 09:24:50
--- 分析対象テーブル数: 2
--- 基づいた分析: LLMによるLiquid Clustering分析
diff --git a/liquid_clustering_implementation_20250815_092456.sql b/liquid_clustering_implementation_20250815_092456.sql
deleted file mode 100644
index dc6f716..0000000
--- a/liquid_clustering_implementation_20250815_092456.sql
+++ /dev/null
@@ -1,120 +0,0 @@
--- =====================================================
--- Liquid Clustering 実装SQL例
--- 生成日時: 2025-08-15 09:24:56
--- =====================================================
-
--- 【重要】
--- 以下のSQL例は分析結果に基づく推奨事項です。
--- 実際の実装前に、テーブル構造やデータ特性を確認してください。
-
-
--- =====================================================
--- テーブル: tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo
--- 現在のクラスタリングキー: cs_sold_date_sk
--- =====================================================
-
--- 既存テーブルにLiquid Clusteringを適用する場合:
--- ALTER TABLE tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo CLUSTER BY (column1, column2, column3, column4);
-
--- 新規テーブル作成時にLiquid Clusteringを設定する場合:
--- CREATE TABLE tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo_clustered
--- CLUSTER BY (column1, column2, column3, column4)
--- AS SELECT * FROM tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo;
-
--- Delta Live Tablesでの設定例:
--- @dlt.table(
---   cluster_by=["column1", "column2", "column3", "column4"]
--- )
--- def catalog_sales_demo_clustered():
---   return spark.table("tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo")
-
--- クラスタリング状況の確認:
--- DESCRIBE DETAIL tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo;
-
--- クラスタリング統計の確認:
--- ANALYZE TABLE tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo COMPUTE STATISTICS FOR ALL COLUMNS;
-
-
--- =====================================================
--- テーブル: tpcds.tpcds_sf10000_delta_lc.date_dim
--- 現在のクラスタリングキー: d_date_sk
--- =====================================================
-
--- 既存テーブルにLiquid Clusteringを適用する場合:
--- ALTER TABLE tpcds.tpcds_sf10000_delta_lc.date_dim CLUSTER BY (column1, column2, column3, column4);
-
--- 新規テーブル作成時にLiquid Clusteringを設定する場合:
--- CREATE TABLE tpcds.tpcds_sf10000_delta_lc.date_dim_clustered
--- CLUSTER BY (column1, column2, column3, column4)
--- AS SELECT * FROM tpcds.tpcds_sf10000_delta_lc.date_dim;
-
--- Delta Live Tablesでの設定例:
--- @dlt.table(
---   cluster_by=["column1", "column2", "column3", "column4"]
--- )
--- def date_dim_clustered():
---   return spark.table("tpcds.tpcds_sf10000_delta_lc.date_dim")
-
--- クラスタリング状況の確認:
--- DESCRIBE DETAIL tpcds.tpcds_sf10000_delta_lc.date_dim;
-
--- クラスタリング統計の確認:
--- ANALYZE TABLE tpcds.tpcds_sf10000_delta_lc.date_dim COMPUTE STATISTICS FOR ALL COLUMNS;
-
-
--- =====================================================
--- 一般的なLiquid Clustering実装パターン
--- =====================================================
-
--- パターン1: フィルター頻度の高いカラムを優先
--- 推奨順序: 1) フィルター条件カラム 2) JOIN条件カラム 3) GROUP BYカラム
-
--- パターン2: カーディナリティを考慮した順序
--- 低カーディナリティ → 高カーディナリティの順で配置
-
--- パターン3: データアクセスパターンに基づく配置
--- よく一緒に使用されるカラムを近い位置に配置
-
--- =====================================================
--- 実装後のパフォーマンス検証SQL
--- =====================================================
-
--- 1. クエリ実行計画の確認
--- EXPLAIN SELECT ... FROM table_name WHERE ...;
-
--- 2. ファイルスキップ統計の確認
--- SELECT * FROM table_name WHERE filter_column = 'value';
--- -- SQLプロファイラーでファイルスキップ数を確認
-
--- 3. データ配置の確認
--- SELECT 
---   file_path,
---   count(*) as row_count,
---   min(cluster_column1) as min_val,
---   max(cluster_column1) as max_val
--- FROM table_name
--- GROUP BY file_path
--- ORDER BY file_path;
-
--- =====================================================
--- 注意事項
--- =====================================================
-
--- 1. Liquid Clusteringは最大4カラムまで指定可能
--- 2. パーティショニングとは併用不可
--- 3. 既存のZORDER BYは自動的に無効化される
--- 4. クラスタリングの効果は時間とともに向上する（OPTIMIZE実行で最適化）
--- 5. 定期的なOPTIMIZE実行を推奨
--- 6. **重要**: カラムの指定順序はパフォーマンスに影響しません
---    * CLUSTER BY (col1, col2, col3) と CLUSTER BY (col3, col1, col2) は同等
---    * 従来のパーティショニングやZ-ORDERとは異なる重要な特性
-
--- OPTIMIZE実行例:
--- OPTIMIZE table_name;
-
--- =====================================================
--- 生成情報
--- =====================================================
--- 生成日時: 2025-08-15 09:24:56
--- 分析対象テーブル数: 2
--- 基づいた分析: LLMによるLiquid Clustering分析
diff --git a/liquid_clustering_implementation_20250815_092500.sql b/liquid_clustering_implementation_20250815_092500.sql
deleted file mode 100644
index b2cc9b6..0000000
--- a/liquid_clustering_implementation_20250815_092500.sql
+++ /dev/null
@@ -1,120 +0,0 @@
--- =====================================================
--- Liquid Clustering 実装SQL例
--- 生成日時: 2025-08-15 09:25:00
--- =====================================================
-
--- 【重要】
--- 以下のSQL例は分析結果に基づく推奨事項です。
--- 実際の実装前に、テーブル構造やデータ特性を確認してください。
-
-
--- =====================================================
--- テーブル: tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo
--- 現在のクラスタリングキー: cs_sold_date_sk
--- =====================================================
-
--- 既存テーブルにLiquid Clusteringを適用する場合:
--- ALTER TABLE tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo CLUSTER BY (column1, column2, column3, column4);
-
--- 新規テーブル作成時にLiquid Clusteringを設定する場合:
--- CREATE TABLE tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo_clustered
--- CLUSTER BY (column1, column2, column3, column4)
--- AS SELECT * FROM tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo;
-
--- Delta Live Tablesでの設定例:
--- @dlt.table(
---   cluster_by=["column1", "column2", "column3", "column4"]
--- )
--- def catalog_sales_demo_clustered():
---   return spark.table("tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo")
-
--- クラスタリング状況の確認:
--- DESCRIBE DETAIL tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo;
-
--- クラスタリング統計の確認:
--- ANALYZE TABLE tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo COMPUTE STATISTICS FOR ALL COLUMNS;
-
-
--- =====================================================
--- テーブル: tpcds.tpcds_sf10000_delta_lc.date_dim
--- 現在のクラスタリングキー: d_date_sk
--- =====================================================
-
--- 既存テーブルにLiquid Clusteringを適用する場合:
--- ALTER TABLE tpcds.tpcds_sf10000_delta_lc.date_dim CLUSTER BY (column1, column2, column3, column4);
-
--- 新規テーブル作成時にLiquid Clusteringを設定する場合:
--- CREATE TABLE tpcds.tpcds_sf10000_delta_lc.date_dim_clustered
--- CLUSTER BY (column1, column2, column3, column4)
--- AS SELECT * FROM tpcds.tpcds_sf10000_delta_lc.date_dim;
-
--- Delta Live Tablesでの設定例:
--- @dlt.table(
---   cluster_by=["column1", "column2", "column3", "column4"]
--- )
--- def date_dim_clustered():
---   return spark.table("tpcds.tpcds_sf10000_delta_lc.date_dim")
-
--- クラスタリング状況の確認:
--- DESCRIBE DETAIL tpcds.tpcds_sf10000_delta_lc.date_dim;
-
--- クラスタリング統計の確認:
--- ANALYZE TABLE tpcds.tpcds_sf10000_delta_lc.date_dim COMPUTE STATISTICS FOR ALL COLUMNS;
-
-
--- =====================================================
--- 一般的なLiquid Clustering実装パターン
--- =====================================================
-
--- パターン1: フィルター頻度の高いカラムを優先
--- 推奨順序: 1) フィルター条件カラム 2) JOIN条件カラム 3) GROUP BYカラム
-
--- パターン2: カーディナリティを考慮した順序
--- 低カーディナリティ → 高カーディナリティの順で配置
-
--- パターン3: データアクセスパターンに基づく配置
--- よく一緒に使用されるカラムを近い位置に配置
-
--- =====================================================
--- 実装後のパフォーマンス検証SQL
--- =====================================================
-
--- 1. クエリ実行計画の確認
--- EXPLAIN SELECT ... FROM table_name WHERE ...;
-
--- 2. ファイルスキップ統計の確認
--- SELECT * FROM table_name WHERE filter_column = 'value';
--- -- SQLプロファイラーでファイルスキップ数を確認
-
--- 3. データ配置の確認
--- SELECT 
---   file_path,
---   count(*) as row_count,
---   min(cluster_column1) as min_val,
---   max(cluster_column1) as max_val
--- FROM table_name
--- GROUP BY file_path
--- ORDER BY file_path;
-
--- =====================================================
--- 注意事項
--- =====================================================
-
--- 1. Liquid Clusteringは最大4カラムまで指定可能
--- 2. パーティショニングとは併用不可
--- 3. 既存のZORDER BYは自動的に無効化される
--- 4. クラスタリングの効果は時間とともに向上する（OPTIMIZE実行で最適化）
--- 5. 定期的なOPTIMIZE実行を推奨
--- 6. **重要**: カラムの指定順序はパフォーマンスに影響しません
---    * CLUSTER BY (col1, col2, col3) と CLUSTER BY (col3, col1, col2) は同等
---    * 従来のパーティショニングやZ-ORDERとは異なる重要な特性
-
--- OPTIMIZE実行例:
--- OPTIMIZE table_name;
-
--- =====================================================
--- 生成情報
--- =====================================================
--- 生成日時: 2025-08-15 09:25:00
--- 分析対象テーブル数: 2
--- 基づいた分析: LLMによるLiquid Clustering分析
diff --git a/output_enhanced_shuffle_analysis_jp_20250815_092450.md b/output_enhanced_shuffle_analysis_jp_20250815_092450.md
deleted file mode 100644
index fc88bc3..0000000
--- a/output_enhanced_shuffle_analysis_jp_20250815_092450.md
+++ /dev/null
@@ -1,89 +0,0 @@
-
-================================================================================
-🔧 Enhanced SHUFFLE操作最適化分析レポート
-================================================================================
-📊 基準: メモリ/パーティション ≤ 512MB
-================================================================================
-
-📊 全体サマリー:
-  ・Shuffle操作数: 3
-  ・最適化が必要な操作: 1
-  ・総メモリ使用量: 405.66 GB
-  ・平均メモリ/パーティション: 6293.8 MB
-  ・最適化必要性: はい
-
-🎯 Shuffle効率性スコア: 🟡 66.7%
-
-🔍 個別Shuffle操作分析:
-
-1. Shuffle (Node ID: 6309)
-   🚨 優先度: HIGH
-   📊 パーティション数: 64
-   🧠 ピークメモリ: 405.39 GB
-   ⚡ メモリ/パーティション: 6486.2 MB 🔥 危険レベル
-   ⏱️ 実行時間: 3698.2 秒
-   📈 処理行数: 5,467,120,697
-   🎯 効率性: ❌ 非効率
-
-   💡 推奨事項:
-     - 🚨 非常に高いメモリ使用量 (6486MB/パーティション): パーティション数を810以上に増加するか、クラスターサイズを拡張してください
-     - 🖥️ クラスター拡張: より多くのワーカーノードまたは高メモリインスタンスの使用を検討
-     - 🔧 Liquid Clusteringの実装により、Shuffle操作の削減を検討 (現在のメモリ使用量: 405.4GB)
-     - ⏱️ 実行時間が長い (3698.2秒): データ分散戦略の見直しを推奨
-     - 📊 大量データ処理 (5,467,120,697行): ブロードキャストJOINや事前集約の活用を検討
-     - 🔧 SQLクエリで発生している場合はREPARTITONヒントもしくはREPARTITON_BY_RANGEヒント(Window関数使用時)を適切に設定してください
-
-2. Shuffle (Node ID: 6317)
-   💡 優先度: LOW
-   📊 パーティション数: 1
-   🧠 ピークメモリ: 0.26 GB
-   ⚡ メモリ/パーティション: 268.0 MB
-   ⏱️ 実行時間: 0.3 秒
-   📈 処理行数: 640,064
-   🎯 効率性: ✅ 効率的
-
-3. Shuffle (Node ID: 6226)
-   💡 優先度: LOW
-   📊 パーティション数: 1
-   🧠 ピークメモリ: 0.0 GB
-   ⚡ メモリ/パーティション: 4.8 MB
-   ⏱️ 実行時間: 0.8 秒
-   📈 処理行数: 160,708
-   🎯 効率性: ✅ 効率的
-
-🎯 全体最適化推奨事項:
-
-  🔧 1/3 のShuffle操作で最適化が必要 (効率性: 66.7%)
-  💎 Liquid Clusteringの実装により根本的なShuffle削減を推奨 (最も効果的な長期解決策)
-  ⚙️ 適切なパーティション数への調整でメモリ効率を改善 (目標: ≤512MB/パーティション)
-  🖥️ クラスターサイズの拡張でメモリ圧迫を軽減 (高優先度ケースで推奨)
-
-📋 実装手順 (優先度順):
-
-1️⃣ 緊急対策 (高優先度ノード向け):
-   - クラスターサイズの拡張 (ワーカーノード数増加)
-   - 高メモリインスタンスタイプへの変更
-   - spark.sql.adaptive.coalescePartitions.maxBatchSize の調整
-
-2️⃣ 短期対策 (即座に実行可能):
-   - spark.sql.adaptive.coalescePartitions.enabled = true
-   - spark.sql.adaptive.skewJoin.enabled = true
-   - spark.sql.adaptive.advisoryPartitionSizeInBytes の調整
-   - 目標: 512MB/パーティション以下
-
-3️⃣ 中期対策 (計画的実装):
-   - パーティション数の明示的指定 (.repartition())
-   - JOIN戦略の最適化 (ブロードキャストJOINの活用)
-   - データ分散戦略の見直し
-
-4️⃣ 長期対策 (根本的解決):
-   - Liquid Clusteringの実装
-   - テーブル設計の最適化
-   - ワークロード分離の検討
-
-⚙️ 推奨Sparkパラメータ:
-
-spark.sql.adaptive.advisoryPartitionSizeInBytes = 268435456
-spark.sql.adaptive.coalescePartitions.minPartitionNum = 1
-spark.sql.adaptive.coalescePartitions.maxBatchSize = 100
-spark.sql.adaptive.skewJoin.skewedPartitionThresholdInBytes = 268435456
diff --git a/output_enhanced_shuffle_analysis_jp_20250815_092456.md b/output_enhanced_shuffle_analysis_jp_20250815_092456.md
deleted file mode 100644
index fc88bc3..0000000
--- a/output_enhanced_shuffle_analysis_jp_20250815_092456.md
+++ /dev/null
@@ -1,89 +0,0 @@
-
-================================================================================
-🔧 Enhanced SHUFFLE操作最適化分析レポート
-================================================================================
-📊 基準: メモリ/パーティション ≤ 512MB
-================================================================================
-
-📊 全体サマリー:
-  ・Shuffle操作数: 3
-  ・最適化が必要な操作: 1
-  ・総メモリ使用量: 405.66 GB
-  ・平均メモリ/パーティション: 6293.8 MB
-  ・最適化必要性: はい
-
-🎯 Shuffle効率性スコア: 🟡 66.7%
-
-🔍 個別Shuffle操作分析:
-
-1. Shuffle (Node ID: 6309)
-   🚨 優先度: HIGH
-   📊 パーティション数: 64
-   🧠 ピークメモリ: 405.39 GB
-   ⚡ メモリ/パーティション: 6486.2 MB 🔥 危険レベル
-   ⏱️ 実行時間: 3698.2 秒
-   📈 処理行数: 5,467,120,697
-   🎯 効率性: ❌ 非効率
-
-   💡 推奨事項:
-     - 🚨 非常に高いメモリ使用量 (6486MB/パーティション): パーティション数を810以上に増加するか、クラスターサイズを拡張してください
-     - 🖥️ クラスター拡張: より多くのワーカーノードまたは高メモリインスタンスの使用を検討
-     - 🔧 Liquid Clusteringの実装により、Shuffle操作の削減を検討 (現在のメモリ使用量: 405.4GB)
-     - ⏱️ 実行時間が長い (3698.2秒): データ分散戦略の見直しを推奨
-     - 📊 大量データ処理 (5,467,120,697行): ブロードキャストJOINや事前集約の活用を検討
-     - 🔧 SQLクエリで発生している場合はREPARTITONヒントもしくはREPARTITON_BY_RANGEヒント(Window関数使用時)を適切に設定してください
-
-2. Shuffle (Node ID: 6317)
-   💡 優先度: LOW
-   📊 パーティション数: 1
-   🧠 ピークメモリ: 0.26 GB
-   ⚡ メモリ/パーティション: 268.0 MB
-   ⏱️ 実行時間: 0.3 秒
-   📈 処理行数: 640,064
-   🎯 効率性: ✅ 効率的
-
-3. Shuffle (Node ID: 6226)
-   💡 優先度: LOW
-   📊 パーティション数: 1
-   🧠 ピークメモリ: 0.0 GB
-   ⚡ メモリ/パーティション: 4.8 MB
-   ⏱️ 実行時間: 0.8 秒
-   📈 処理行数: 160,708
-   🎯 効率性: ✅ 効率的
-
-🎯 全体最適化推奨事項:
-
-  🔧 1/3 のShuffle操作で最適化が必要 (効率性: 66.7%)
-  💎 Liquid Clusteringの実装により根本的なShuffle削減を推奨 (最も効果的な長期解決策)
-  ⚙️ 適切なパーティション数への調整でメモリ効率を改善 (目標: ≤512MB/パーティション)
-  🖥️ クラスターサイズの拡張でメモリ圧迫を軽減 (高優先度ケースで推奨)
-
-📋 実装手順 (優先度順):
-
-1️⃣ 緊急対策 (高優先度ノード向け):
-   - クラスターサイズの拡張 (ワーカーノード数増加)
-   - 高メモリインスタンスタイプへの変更
-   - spark.sql.adaptive.coalescePartitions.maxBatchSize の調整
-
-2️⃣ 短期対策 (即座に実行可能):
-   - spark.sql.adaptive.coalescePartitions.enabled = true
-   - spark.sql.adaptive.skewJoin.enabled = true
-   - spark.sql.adaptive.advisoryPartitionSizeInBytes の調整
-   - 目標: 512MB/パーティション以下
-
-3️⃣ 中期対策 (計画的実装):
-   - パーティション数の明示的指定 (.repartition())
-   - JOIN戦略の最適化 (ブロードキャストJOINの活用)
-   - データ分散戦略の見直し
-
-4️⃣ 長期対策 (根本的解決):
-   - Liquid Clusteringの実装
-   - テーブル設計の最適化
-   - ワークロード分離の検討
-
-⚙️ 推奨Sparkパラメータ:
-
-spark.sql.adaptive.advisoryPartitionSizeInBytes = 268435456
-spark.sql.adaptive.coalescePartitions.minPartitionNum = 1
-spark.sql.adaptive.coalescePartitions.maxBatchSize = 100
-spark.sql.adaptive.skewJoin.skewedPartitionThresholdInBytes = 268435456
diff --git a/output_enhanced_shuffle_analysis_jp_20250815_092500.md b/output_enhanced_shuffle_analysis_jp_20250815_092500.md
deleted file mode 100644
index fc88bc3..0000000
--- a/output_enhanced_shuffle_analysis_jp_20250815_092500.md
+++ /dev/null
@@ -1,89 +0,0 @@
-
-================================================================================
-🔧 Enhanced SHUFFLE操作最適化分析レポート
-================================================================================
-📊 基準: メモリ/パーティション ≤ 512MB
-================================================================================
-
-📊 全体サマリー:
-  ・Shuffle操作数: 3
-  ・最適化が必要な操作: 1
-  ・総メモリ使用量: 405.66 GB
-  ・平均メモリ/パーティション: 6293.8 MB
-  ・最適化必要性: はい
-
-🎯 Shuffle効率性スコア: 🟡 66.7%
-
-🔍 個別Shuffle操作分析:
-
-1. Shuffle (Node ID: 6309)
-   🚨 優先度: HIGH
-   📊 パーティション数: 64
-   🧠 ピークメモリ: 405.39 GB
-   ⚡ メモリ/パーティション: 6486.2 MB 🔥 危険レベル
-   ⏱️ 実行時間: 3698.2 秒
-   📈 処理行数: 5,467,120,697
-   🎯 効率性: ❌ 非効率
-
-   💡 推奨事項:
-     - 🚨 非常に高いメモリ使用量 (6486MB/パーティション): パーティション数を810以上に増加するか、クラスターサイズを拡張してください
-     - 🖥️ クラスター拡張: より多くのワーカーノードまたは高メモリインスタンスの使用を検討
-     - 🔧 Liquid Clusteringの実装により、Shuffle操作の削減を検討 (現在のメモリ使用量: 405.4GB)
-     - ⏱️ 実行時間が長い (3698.2秒): データ分散戦略の見直しを推奨
-     - 📊 大量データ処理 (5,467,120,697行): ブロードキャストJOINや事前集約の活用を検討
-     - 🔧 SQLクエリで発生している場合はREPARTITONヒントもしくはREPARTITON_BY_RANGEヒント(Window関数使用時)を適切に設定してください
-
-2. Shuffle (Node ID: 6317)
-   💡 優先度: LOW
-   📊 パーティション数: 1
-   🧠 ピークメモリ: 0.26 GB
-   ⚡ メモリ/パーティション: 268.0 MB
-   ⏱️ 実行時間: 0.3 秒
-   📈 処理行数: 640,064
-   🎯 効率性: ✅ 効率的
-
-3. Shuffle (Node ID: 6226)
-   💡 優先度: LOW
-   📊 パーティション数: 1
-   🧠 ピークメモリ: 0.0 GB
-   ⚡ メモリ/パーティション: 4.8 MB
-   ⏱️ 実行時間: 0.8 秒
-   📈 処理行数: 160,708
-   🎯 効率性: ✅ 効率的
-
-🎯 全体最適化推奨事項:
-
-  🔧 1/3 のShuffle操作で最適化が必要 (効率性: 66.7%)
-  💎 Liquid Clusteringの実装により根本的なShuffle削減を推奨 (最も効果的な長期解決策)
-  ⚙️ 適切なパーティション数への調整でメモリ効率を改善 (目標: ≤512MB/パーティション)
-  🖥️ クラスターサイズの拡張でメモリ圧迫を軽減 (高優先度ケースで推奨)
-
-📋 実装手順 (優先度順):
-
-1️⃣ 緊急対策 (高優先度ノード向け):
-   - クラスターサイズの拡張 (ワーカーノード数増加)
-   - 高メモリインスタンスタイプへの変更
-   - spark.sql.adaptive.coalescePartitions.maxBatchSize の調整
-
-2️⃣ 短期対策 (即座に実行可能):
-   - spark.sql.adaptive.coalescePartitions.enabled = true
-   - spark.sql.adaptive.skewJoin.enabled = true
-   - spark.sql.adaptive.advisoryPartitionSizeInBytes の調整
-   - 目標: 512MB/パーティション以下
-
-3️⃣ 中期対策 (計画的実装):
-   - パーティション数の明示的指定 (.repartition())
-   - JOIN戦略の最適化 (ブロードキャストJOINの活用)
-   - データ分散戦略の見直し
-
-4️⃣ 長期対策 (根本的解決):
-   - Liquid Clusteringの実装
-   - テーブル設計の最適化
-   - ワークロード分離の検討
-
-⚙️ 推奨Sparkパラメータ:
-
-spark.sql.adaptive.advisoryPartitionSizeInBytes = 268435456
-spark.sql.adaptive.coalescePartitions.minPartitionNum = 1
-spark.sql.adaptive.coalescePartitions.maxBatchSize = 100
-spark.sql.adaptive.skewJoin.skewedPartitionThresholdInBytes = 268435456
diff --git a/output_final_report_en_20250815-111615.md b/output_final_report_en_20250815-111615.md
deleted file mode 100644
index 4fc9f59..0000000
--- a/output_final_report_en_20250815-111615.md
+++ /dev/null
@@ -1,255 +0,0 @@
-# 📊 SQL Optimization Report
-
-**Query ID**: 01f078e6-dc5c-1a82-902a-652166ae2162  
-**Report Generation Time**: 2025-08-15 11:14:04
-
-## 🎯 Executive Summary
-
-This query is experiencing significant performance issues primarily due to inefficient shuffle operations, poor cache utilization, and suboptimal data filtering. The execution time of 253.6 seconds can be reduced by approximately 45% through table optimization and query tuning.
-
-## 🔍 Performance Analysis
-
-### Key Performance Indicators
-
-| Metric | Value | Status |
-|--------|-------|--------|
-| Execution Time | 253.6s | ⚠️ Needs Improvement |
-| Data Read | 159.08GB | ⚠️ Large Volume |
-| Photon Utilization | 99.4% | ✅ Good |
-| Cache Efficiency | 0.0% | ⚠️ Needs Improvement |
-| Filter Rate | 5.1% | ⚠️ Check Filter Conditions |
-| Shuffle Impact | 50.2% impact | ❌ Serious Optimization Needed |
-| Spill Occurred | No | ✅ Good |
-| Skew Detection | Not detected | ✅ Good |
-
-### Primary Bottlenecks
-
-1. **Shuffle Operations**: Consuming 50.2% of execution time
-2. **Inefficient Cache Usage**: 0% cache hit rate
-3. **Poor Data Filtering**: Reading excessive data
-
-## 🐌 Time-Consuming Operations
-
-### Top 5 Performance Bottlenecks
-
-1. **Photon Shuffle Exchange** (50.2% of total time)
-   - Execution time: 3,698,225 ms (3698.2 sec)
-   - Peak memory: 415120.0 MB
-   - 🔧 Parallelism: Sink - Tasks total: 1333 | Source - Tasks total: 64
-   - Shuffle attribute: tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo.cs_bill_customer_sk
-   - Node ID: 6309
-
-2. **Photon Grouping Aggregate** (26.6% of total time)
-   - Execution time: 1,962,151 ms (1962.2 sec)
-   - Peak memory: 491298.0 MB
-   - 🔧 Parallelism: Tasks total: 1333
-   - Node ID: 6232
-
-3. **Photon Data Source Scan** (14.3% of total time)
-   - Execution time: 1,053,662 ms (1053.7 sec)
-   - Peak memory: 22755.6 MB
-   - 🔧 Parallelism: Tasks total: 1333
-   - Filter rate: 87.6% (read: 1279.14GB, actual: 159.06GB)
-   - Current clustering key: cs_sold_date_sk
-   - Node ID: 6224
-
-4. **Photon Grouping Aggregate** (7.0% of total time)
-   - Execution time: 512,669 ms (512.7 sec)
-   - Peak memory: 5504.0 MB
-   - 🔧 Parallelism: Tasks total: 64
-   - Node ID: 6311
-
-5. **Photon Left Semi Join** (1.1% of total time)
-   - Execution time: 84,455 ms (84.5 sec)
-   - Peak memory: 28.0 MB
-   - 🔧 Parallelism: Tasks total: 1333
-   - Node ID: 6228
-
-## 📋 Table Optimization Recommendations
-
-### 1. catalog_sales_demo Table (High Priority)
-
-**Table Information:**
-- Size: 1279.14GB
-- Current clustering key: cs_sold_date_sk
-- Recommended clustering columns: cs_sold_date_sk, cs_bill_customer_sk
-- Filter rate: 5.1% (read: 1279.14GB, pruned: 1214.4GB)
-
-**Implementation SQL:**
-```sql
-ALTER TABLE tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo 
-CLUSTER BY (cs_sold_date_sk, cs_bill_customer_sk);
-OPTIMIZE tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo FULL;
-```
-
-**Rationale:**
-- Large table size (1279.14GB) makes clustering optimization highly beneficial
-- cs_sold_date_sk:
-  - Used in filter conditions (IS NOT NULL)
-  - Used as JOIN key
-  - Already configured as current clustering key
-- cs_bill_customer_sk:
-  - Used twice in GROUP BY operations
-  - Central column for data aggregation
-- Note: In Liquid Clustering, key order doesn't affect node-level data locality
-
-**Expected Benefits:**
-- Query execution time reduction: 30-40% (253.6s → ~150-180s)
-- Shuffle impact reduction: from 50.2% to ~30%
-- Improved GROUP BY operation efficiency
-
-### 2. date_dim Table (Not Recommended)
-
-**Table Information:**
-- Size: 0.00GB
-- Current clustering key: d_date_sk
-- Filter rate: 54.4% (read: 0.00GB, actual: 0.00GB)
-
-**Alternative Recommendations:**
-```sql
--- ❌ Liquid Clustering not recommended due to small table size
--- 💡 Alternative: CACHE TABLE tpcds.tpcds_sf10000_delta_lc.date_dim;
--- 💡 Or: OPTIMIZE tpcds.tpcds_sf10000_delta_lc.date_dim;
-```
-
-**Rationale:**
-- Extremely small table size makes clustering ineffective
-- For small tables, memory caching is more efficient
-- Current key (d_date_sk) already covers JOIN and filter conditions
-
-## 🚀 Query Optimization Results
-
-### Optimization Process
-
-**Trial History:**
-- 2 optimization attempts executed
-- Final selection: Original query (no improvements achieved)
-- Reason: Optimization trials did not yield significant improvements
-
-**Key Issues:**
-- Shuffle processing bottleneck
-- Low cache hit rate
-
-**Recommended Query:**
-```sql
-USE CATALOG tpcds;
-USE SCHEMA tpcds_sf1000_delta_lc;
--- 集計クエリ
-SELECT
-cs_bill_customer_sk,
-AVG(cs_ext_sales_price) AS avg_cs_ext_sales_price,
-MIN(cs_ext_sales_price) AS min_cs_ext_sales_price,
-MAX(cs_ext_sales_price) AS max_cs_ext_sales_price,
-COUNT(cs_ext_sales_price) AS count_cs_ext_sales_price,
-AVG(cs_net_profit) AS avg_cs_net_profit,
-MIN(cs_net_profit) AS min_cs_net_profit,
-MAX(cs_net_profit) AS max_cs_net_profit,
-COUNT(cs_net_profit) AS count_cs_net_profit
-FROM
-tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo
-WHERE
-cs_sold_date_sk in (
-select
-d_date_sk
-from
-tpcds.tpcds_sf10000_delta_lc.date_dim
-where
-d_date >= '1990-01-02'
-)
-GROUP BY
-cs_bill_customer_sk
-ORDER BY
-cs_bill_customer_sk;
-```
-
-### Expected Performance Improvement
-
-**Anticipated Improvements:**
-1. Shuffle optimization: 20-60% execution time reduction
-2. Cache efficiency: 30-70% read time reduction
-3. Filter efficiency: 40-90% data read volume reduction
-
-**Overall improvement:** Execution time reduction from 253,607ms to ~139,484ms (45% improvement)
-
-## 📋 Additional Table Analysis
-
-### catalog_sales Table
-
-**Basic Information:**
-- Table size: 1220.35GB
-- Current clustering key: cs_item_sk, cs_sold_date_sk
-- Recommended clustering columns: cs_bill_customer_sk, cs_item_sk, cs_sold_date_sk
-
-**Implementation SQL:**
-```sql
-ALTER TABLE tpcds.tpcds_sf10000_delta_lc.catalog_sales 
-CLUSTER BY (cs_bill_customer_sk, cs_item_sk, cs_sold_date_sk);
-OPTIMIZE tpcds.tpcds_sf10000_delta_lc.catalog_sales FULL;
-```
-
-**Expected Benefits:**
-- 30-40% execution time reduction
-- Reduced shuffle operations and spills
-
-### 💡 Liquid Clustering Key Selection Guidelines
-
-**Key Selection Principles:**
-- Focus on read optimization via data skipping on filter columns
-- Prioritize columns frequently used for filtering
-
-**GROUP BY Key Consideration Conditions:**
-1. When filter columns also appear in GROUP BY
-2. When intermediate data volume reduction is expected
-3. When keys have low to medium cardinality with minimal skew
-
-**Practical Recommendation:**
-If the above conditions aren't met, prioritize filter columns
-
-## 🔧 Enhanced Shuffle Operations Optimization Analysis
-
-📊 Threshold: Memory per Partition ≤ 512MB
-
-📊 Overall Summary:
-  • Number of Shuffle Operations: 3
-  • Operations Requiring Optimization: 1
-  • Total Memory Usage: 405.66 GB
-  • Average Memory per Partition: 6293.8 MB
-  • Optimization Required: Yes
-
-🎯 Shuffle Efficiency Score: 🟡 66.7%
-
-## 🔍 Execution Plan Analysis
-
-```
-== Physical Plan ==
-AdaptiveSparkPlan isFinalPlan=false
-+- == Initial Plan ==
-   ColumnarToRow
-   +- PhotonResultStage
-      +- PhotonSort [cs_bill_customer_sk#50084 ASC NULLS FIRST]
-         +- PhotonShuffleExchangeSource
-            +- PhotonShuffleMapStage
-               +- PhotonShuffleExchangeSink rangepartitioning(cs_bill_customer_sk#50084 ASC NULLS FIRST, 16)
-                  +- PhotonGroupingAgg(keys=[cs_bill_customer_sk#50084], functions=[avg(UnscaledValue(cs_ext_sales_price#50104)), min(cs_ext_sales_price#50104), max(cs_ext_sales_price#50104), count(cs_ext_sales_price#50104), avg(UnscaledValue(cs_net_profit#50114)), min(cs_net_profit#50114), max(cs_net_profit#50114), count(cs_net_profit#50114)])
-                     +- PhotonShuffleExchangeSource
-                        +- PhotonShuffleMapStage
-                           +- PhotonShuffleExchangeSink hashpartitioning(cs_bill_customer_sk#50084, 2666)
-                              +- PhotonProject [cs_bill_customer_sk#50084, cs_ext_sales_price#50104, cs_net_profit#50114]
-                                 +- PhotonBroadcastHashJoin [cs_sold_date_sk#50081], [d_date_sk#50115], Inner, BuildRight, false, true
-                                    :- PhotonScan parquet tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo[cs_sold_date_sk#50081,cs_bill_customer_sk#50084,cs_ext_sales_price#50104,cs_net_profit#50114] DataFilters: [isnotnull(cs_sold_date_sk#50081), dynamicpruning#50168 50166], DictionaryFilters: [], Format: parquet, Location: PreparedDeltaFileIndex(1 paths)[s3://e2-demo-tokyo-uc/6ba60eaa-3923-4fda-bed0-42216b8451e0/tables..., OptionalDataFilters: [hashedrelationcontains(cs_sold_date_sk#50081)], PartitionFilters: [], ReadSchema: struct<cs_sold_date_sk:int,cs_bill_customer_sk:int,cs_ext_sales_price:decimal(7,2),cs_net_profit:..., RequiredDataFilters: [isnotnull(cs_sold_date_sk#50081)]
-                                    +- PhotonShuffleExchangeSource
-                                       +- PhotonShuffleMapStage
-                                          +- PhotonShuffleExchangeSink SinglePartition
-                                             +- PhotonProject [d_date_sk#50115]
-                                                +- PhotonScan parquet tpcds.tpcds_sf10000_delta_lc.date_dim[d_date_sk#50115,d_date#50117] DataFilters: [isnotnull(d_date#50117), isnotnull(d_date_sk#50115), (d_date#50117 >= 1990-01-02)], DictionaryFilters: [(d_date#50117 >= 1990-01-02)], Format: parquet, Location: PreparedDeltaFileIndex(1 paths)[s3://e2-demo-tokyo-uc/wanyu/tpcds-2.13/tpcds_sf10000_delta_lc/dat..., OptionalDataFilters: [], PartitionFilters: [], ReadSchema: struct<d_date_sk:int,d_date:date>, RequiredDataFilters: [isnotnull(d_date#50117), isnotnull(d_date_sk#50115), (d_date#50117 >= 1990-01-02)]
-
-
-== Photon Explanation ==
-The query is fully supported by Photon.
-== Optimizer Statistics (table names per statistics state) ==
-  missing = 
-  partial = 
-  full    = catalog_sales_demo, date_dim
-```
-
-*Report generated at: 2025-08-15 11:14:04*
\ No newline at end of file
diff --git a/output_original_query_20250815-092450.sql b/output_original_query_20250815-092450.sql
deleted file mode 100644
index a55f61e..0000000
--- a/output_original_query_20250815-092450.sql
+++ /dev/null
@@ -1,48 +0,0 @@
--- 📋 オリジナルクエリ（プロファイラーデータから抽出）
--- 抽出日時: 2025-08-15 09:24:50
--- ファイル: .//output_original_query_20250815-092450.sql
--- クエリ文字数: 1,401
-
--- 🗂️ カタログ・スキーマ設定（自動追加）
-USE CATALOG tpcds;
-USE SCHEMA tpcds_sf1000_delta_lc;
-
--- 🔍 オリジナルクエリ
--- 集計クエリ
-SELECT
-  cs_bill_customer_sk,
-  AVG(cs_ext_sales_price) AS avg_cs_ext_sales_price,
-  MIN(cs_ext_sales_price) AS min_cs_ext_sales_price,
-  MAX(cs_ext_sales_price) AS max_cs_ext_sales_price,
-  COUNT(cs_ext_sales_price) AS count_cs_ext_sales_price,
-  AVG(cs_net_profit) AS avg_cs_net_profit,
-  MIN(cs_net_profit) AS min_cs_net_profit,
-  MAX(cs_net_profit) AS max_cs_net_profit,
-  COUNT(cs_net_profit) AS count_cs_net_profit
-  -- AVG(cs_wholesale_cost) AS avg_another_numeric_column1,
-  -- MIN(cs_wholesale_cost) AS min_another_numeric_column1,
-  -- MAX(cs_wholesale_cost) AS max_another_numeric_column1,
-  -- COUNT(cs_wholesale_cost) AS count_another_numeric_column1,
-  -- AVG(cs_list_price) AS avg_another_numeric_column2,
-  -- MIN(cs_list_price) AS min_another_numeric_column2,
-  -- MAX(cs_list_price) AS max_another_numeric_column2,
-  -- COUNT(cs_list_price) AS count_another_numeric_column2,
-  -- AVG(cs_sales_price) AS avg_another_numeric_column3,
-  -- MIN(cs_sales_price) AS min_another_numeric_column3,
-  -- MAX(cs_sales_price) AS max_another_numeric_column3,
-  -- COUNT(cs_sales_price) AS count_another_numeric_column3
-FROM
-  tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo
-WHERE
-  cs_sold_date_sk in (
-    select
-      d_date_sk
-    from
-      tpcds.tpcds_sf10000_delta_lc.date_dim
-    where
-      d_date >= '1990-01-02'
-  )
-GROUP BY
-  cs_bill_customer_sk
-ORDER BY
-  cs_bill_customer_sk
diff --git a/output_original_query_20250815-092456.sql b/output_original_query_20250815-092456.sql
deleted file mode 100644
index 05ef3a7..0000000
--- a/output_original_query_20250815-092456.sql
+++ /dev/null
@@ -1,48 +0,0 @@
--- 📋 オリジナルクエリ（プロファイラーデータから抽出）
--- 抽出日時: 2025-08-15 09:24:56
--- ファイル: .//output_original_query_20250815-092456.sql
--- クエリ文字数: 1,401
-
--- 🗂️ カタログ・スキーマ設定（自動追加）
-USE CATALOG tpcds;
-USE SCHEMA tpcds_sf1000_delta_lc;
-
--- 🔍 オリジナルクエリ
--- 集計クエリ
-SELECT
-  cs_bill_customer_sk,
-  AVG(cs_ext_sales_price) AS avg_cs_ext_sales_price,
-  MIN(cs_ext_sales_price) AS min_cs_ext_sales_price,
-  MAX(cs_ext_sales_price) AS max_cs_ext_sales_price,
-  COUNT(cs_ext_sales_price) AS count_cs_ext_sales_price,
-  AVG(cs_net_profit) AS avg_cs_net_profit,
-  MIN(cs_net_profit) AS min_cs_net_profit,
-  MAX(cs_net_profit) AS max_cs_net_profit,
-  COUNT(cs_net_profit) AS count_cs_net_profit
-  -- AVG(cs_wholesale_cost) AS avg_another_numeric_column1,
-  -- MIN(cs_wholesale_cost) AS min_another_numeric_column1,
-  -- MAX(cs_wholesale_cost) AS max_another_numeric_column1,
-  -- COUNT(cs_wholesale_cost) AS count_another_numeric_column1,
-  -- AVG(cs_list_price) AS avg_another_numeric_column2,
-  -- MIN(cs_list_price) AS min_another_numeric_column2,
-  -- MAX(cs_list_price) AS max_another_numeric_column2,
-  -- COUNT(cs_list_price) AS count_another_numeric_column2,
-  -- AVG(cs_sales_price) AS avg_another_numeric_column3,
-  -- MIN(cs_sales_price) AS min_another_numeric_column3,
-  -- MAX(cs_sales_price) AS max_another_numeric_column3,
-  -- COUNT(cs_sales_price) AS count_another_numeric_column3
-FROM
-  tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo
-WHERE
-  cs_sold_date_sk in (
-    select
-      d_date_sk
-    from
-      tpcds.tpcds_sf10000_delta_lc.date_dim
-    where
-      d_date >= '1990-01-02'
-  )
-GROUP BY
-  cs_bill_customer_sk
-ORDER BY
-  cs_bill_customer_sk
diff --git a/output_original_query_20250815-092500.sql b/output_original_query_20250815-092500.sql
deleted file mode 100644
index 3caffc1..0000000
--- a/output_original_query_20250815-092500.sql
+++ /dev/null
@@ -1,48 +0,0 @@
--- 📋 オリジナルクエリ（プロファイラーデータから抽出）
--- 抽出日時: 2025-08-15 09:25:00
--- ファイル: .//output_original_query_20250815-092500.sql
--- クエリ文字数: 1,401
-
--- 🗂️ カタログ・スキーマ設定（自動追加）
-USE CATALOG tpcds;
-USE SCHEMA tpcds_sf1000_delta_lc;
-
--- 🔍 オリジナルクエリ
--- 集計クエリ
-SELECT
-  cs_bill_customer_sk,
-  AVG(cs_ext_sales_price) AS avg_cs_ext_sales_price,
-  MIN(cs_ext_sales_price) AS min_cs_ext_sales_price,
-  MAX(cs_ext_sales_price) AS max_cs_ext_sales_price,
-  COUNT(cs_ext_sales_price) AS count_cs_ext_sales_price,
-  AVG(cs_net_profit) AS avg_cs_net_profit,
-  MIN(cs_net_profit) AS min_cs_net_profit,
-  MAX(cs_net_profit) AS max_cs_net_profit,
-  COUNT(cs_net_profit) AS count_cs_net_profit
-  -- AVG(cs_wholesale_cost) AS avg_another_numeric_column1,
-  -- MIN(cs_wholesale_cost) AS min_another_numeric_column1,
-  -- MAX(cs_wholesale_cost) AS max_another_numeric_column1,
-  -- COUNT(cs_wholesale_cost) AS count_another_numeric_column1,
-  -- AVG(cs_list_price) AS avg_another_numeric_column2,
-  -- MIN(cs_list_price) AS min_another_numeric_column2,
-  -- MAX(cs_list_price) AS max_another_numeric_column2,
-  -- COUNT(cs_list_price) AS count_another_numeric_column2,
-  -- AVG(cs_sales_price) AS avg_another_numeric_column3,
-  -- MIN(cs_sales_price) AS min_another_numeric_column3,
-  -- MAX(cs_sales_price) AS max_another_numeric_column3,
-  -- COUNT(cs_sales_price) AS count_another_numeric_column3
-FROM
-  tpcds.tpcds_sf10000_delta_lc.catalog_sales_demo
-WHERE
-  cs_sold_date_sk in (
-    select
-      d_date_sk
-    from
-      tpcds.tpcds_sf10000_delta_lc.date_dim
-    where
-      d_date >= '1990-01-02'
-  )
-GROUP BY
-  cs_bill_customer_sk
-ORDER BY
-  cs_bill_customer_sk