DashBoard/spec.md at dd0ae3ee5454ebf85410c9fe5893255ed42c0fe5

Files

beabigegg dd0ae3ee54 feat: Excel 批次查詢新增進階條件功能

- 新增欄位類型偵測：自動識別 Excel 與 Oracle 欄位類型並顯示類型標籤
- 新增 LIKE 模糊查詢：支援包含/開頭/結尾三種模式，上限 100 個關鍵字
- 新增日期範圍篩選：支援起始/結束日期，範圍限制 365 天
- 新增大型資料表效能警告：超過 1000 萬筆時提示使用日期範圍縮小查詢
- 新增 /execute-advanced API 端點整合所有進階條件
- 新增 /table-metadata 端點取得欄位類型資訊
- 新增完整測試套件：76 個測試（單元/整合/E2E）

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

2026-02-04 14:05:19 +08:00

3.7 KiB

Raw Blame History

ADDED Requirements

Requirement: Excel column type detection

The system SHALL analyze Excel column values and detect their data type.

Scenario: Detect date type

WHEN Excel column contains values matching pattern YYYY-MM-DD or YYYY/MM/DD
THEN the system SHALL classify the column as type "date"
AND display type label "日期"

Scenario: Detect datetime type

WHEN Excel column contains values matching pattern YYYY-MM-DD HH:MM or YYYY-MM-DDTHH:MM
THEN the system SHALL classify the column as type "datetime"
AND display type label "日期時間"

Scenario: Detect number type

WHEN Excel column contains values matching pattern ^-?\d+\.?\d*$
THEN the system SHALL classify the column as type "number"
AND display type label "數值"

Scenario: Detect ID type

WHEN Excel column contains values matching pattern ^[A-Z0-9_-]+$ (uppercase alphanumeric with underscore/hyphen)
THEN the system SHALL classify the column as type "id"
AND display type label "識別碼"

Scenario: Default to text type

WHEN Excel column does not match any specific pattern
THEN the system SHALL classify the column as type "text"
AND display type label "文字"

Scenario: Type detection sampling

WHEN system performs type detection
THEN the system SHALL sample the first 100 non-empty values
AND classify based on majority pattern match (>80%)

Requirement: Oracle column metadata retrieval

The system SHALL retrieve column metadata from Oracle database for the selected table.

Scenario: Successful metadata retrieval

WHEN user selects a table
THEN the system SHALL query ALL_TAB_COLUMNS for column information
AND return: COLUMN_NAME, DATA_TYPE, DATA_LENGTH, DATA_PRECISION, DATA_SCALE

Scenario: Metadata query permission denied

WHEN user lacks permission to query ALL_TAB_COLUMNS
THEN the system SHALL fallback to SELECT * FROM {table} WHERE ROWNUM <= 1 method
AND return column names without type information

Requirement: Table metadata API endpoint

The system SHALL provide a new API endpoint /api/excel-query/table-metadata for retrieving enriched table information.

Scenario: Table metadata response

WHEN client calls POST /api/excel-query/table-metadata with {"table_name": "DWH.DW_MES_WIP"}
THEN the system SHALL return:
- columns: array of {name, data_type, is_date, is_number}
- time_field: string or null (from TABLES_CONFIG)
- description: string (from TABLES_CONFIG)
- row_count: number (from TABLES_CONFIG)

Requirement: Column type display in UI

The system SHALL display column type information in the query column selection interface.

Scenario: Oracle column type badges

WHEN user views the query column selection dropdown
THEN each column SHALL display a type badge:
- VARCHAR2/CHAR → "文字"
- NUMBER → "數值"
- DATE/TIMESTAMP → "日期"

Scenario: Excel column type badges

WHEN user views the Excel column selection dropdown
THEN each column SHALL display the detected type badge

Requirement: Column type matching suggestion

The system SHALL suggest compatible column matches between Excel and Oracle columns.

Scenario: Type-compatible suggestion

WHEN user selects an Excel column with detected type "id"
THEN the system SHALL highlight Oracle columns with VARCHAR2 type as "建議"

Scenario: Type-incompatible warning

WHEN user selects an Excel date column but Oracle target column is NUMBER type
THEN the system SHALL display warning: "欄位類型不相符，可能導致查詢結果為空"

3.7 KiB Raw Blame History