admin/mars
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

pRegisterExport

Browse Source
This commit is contained in:
Grzegorz Michalski
2026-02-11 09:53:53 +01:00
parent 3eb39091de
commit 96e3e2f845
2 changed files with 156 additions and 142 deletions
Download Patch File Download Diff File Expand all files Collapse all files

48
MARS_Packages/REL01_ADDITIONS/MARS-835-PREHOOK/new_version/DATA_EXPORTER.pkg
View File

@@ -8,28 +8,27 @@ AS
* which returns documentation text for confluence page (to Copy-Paste it).
**/
-- Package Version Information (Semantic Versioning: MAJOR.MINOR.PATCH)
PACKAGE_VERSION CONSTANT VARCHAR2(10) := '2.7.0';
PACKAGE_BUILD_DATE CONSTANT VARCHAR2(20) := '2026-02-09 20:00:00';
-- Package Version Information
PACKAGE_VERSION CONSTANT VARCHAR2(10) := '2.7.1';
PACKAGE_BUILD_DATE CONSTANT VARCHAR2(20) := '2026-02-11 10:00:00';
PACKAGE_AUTHOR CONSTANT VARCHAR2(100) := 'Grzegorz Michalski';
cgBL CONSTANT VARCHAR2(2) := CHR(13)||CHR(10);
-- Version History (Latest changes first)
VERSION_HISTORY CONSTANT VARCHAR2(4000) :=
'v2.7.0 (2026-02-09): NEW FEATURE - Added pRegisterExport parameter to EXPORT_TABLE_DATA_TO_CSV_BY_DATE. When TRUE, successfully exported files are registered in A_WORKFLOW_HISTORY (one record per YEAR/MONTH) and A_SOURCE_FILE_RECEIVED tables for tracking and audit purposes.' || cgBL ||
'v2.6.3 (2026-01-28): COMPILATION FIX - Resolved ORA-00904 error in EXPORT_PARTITION_PARALLEL. SQLERRM and DBMS_UTILITY.FORMAT_ERROR_BACKTRACE cannot be used directly in SQL UPDATE statements. Now properly assigned to vgMsgTmp variable before UPDATE.' || cgBL ||
'v2.6.2 (2026-01-28): CRITICAL FIX - Race condition when multiple exports run simultaneously. Changed DELETE to filter by age (>24h) instead of deleting all COMPLETED chunks. Prevents concurrent sessions from deleting each other chunks. Session-safe cleanup with TASK_NAME filtering. Enables true parallel execution of multiple export jobs.' || cgBL ||
'v2.6.1 (2026-01-28): Added DELETE_FAILED_EXPORT_FILE procedure to clean up partial/corrupted files before retry. When partition fails mid-export, partial file is deleted before retry to prevent Oracle from creating _1 suffixed duplicates. Ensures clean retry without orphaned files in OCI bucket.' || cgBL ||
'v2.6.0 (2026-01-28): CRITICAL FIX - Added STATUS tracking to A_PARALLEL_EXPORT_CHUNKS table to prevent data duplication on retry. System now restarts ONLY failed partitions instead of re-exporting all data. Added ERROR_MESSAGE and EXPORT_TIMESTAMP columns for better error handling and monitoring. Prevents duplicate file creation when parallel tasks fail (e.g., 22 partitions with 16 threads, 3 failures no longer duplicates 19 successful exports).' || cgBL ||
'v2.5.0 (2026-01-26): Added recorddelimiter parameter with CRLF (CHR(13)||CHR(10)) for CSV exports to ensure Windows-compatible line endings. Improves cross-platform compatibility when CSV files are opened in Windows applications (Notepad, Excel).' || cgBL ||
'v2.4.0 (2026-01-11): Added pTemplateTableName parameter for per-column date format configuration. Implements dynamic query building with TO_CHAR for each date/timestamp column using FILE_MANAGER.GET_DATE_FORMAT. Supports 3-tier hierarchy: column-specific, template DEFAULT, global fallback. Eliminates single dateformat limitation of DBMS_CLOUD.EXPORT_DATA.' || cgBL ||
'v2.3.0 (2025-12-20): Added parallel partition processing using DBMS_PARALLEL_EXECUTE. New pParallelDegree parameter (1-16, default 1) for EXPORT_TABLE_DATA_BY_DATE and EXPORT_TABLE_DATA_TO_CSV_BY_DATE procedures. Each year/month partition processed in separate thread for improved performance.' || cgBL ||
'v2.2.0 (2025-12-19): DRY refactoring - extracted shared helper functions (sanitizeFilename, VALIDATE_TABLE_AND_COLUMNS, GET_PARTITIONS, EXPORT_SINGLE_PARTITION worker procedure). Reduced code duplication by ~400 lines. Prepared architecture for v2.3.0 parallel processing.' || cgBL ||
'v2.1.1 (2025-12-04): Fixed JOIN column reference A_WORKFLOW_HISTORY_KEY -> A_ETL_LOAD_SET_KEY, added consistent column mapping and dynamic column list to EXPORT_TABLE_DATA procedure, enhanced DEBUG logging for all export operations' || cgBL ||
'v2.1.0 (2025-10-22): Added version tracking and PARTITION_YEAR/PARTITION_MONTH support' || cgBL ||
'v2.0.0 (2025-10-01): Separated export functionality from FILE_MANAGER package';
-- Version History (last 3-5 changes)
VERSION_HISTORY CONSTANT VARCHAR2(4000) :=
'v2.7.1 (2026-02-11): AUTO-LOOKUP A_SOURCE_FILE_CONFIG_KEY - Parse pFolderName (format: BUCKET/SOURCE/TABLE_ID) to automatically find config key from A_SOURCE_FILE_CONFIG. Example: ODS/CSDB/CSDB_DEBT_DAILY extracts SOURCE_KEY=CSDB, TABLE_ID=CSDB_DEBT_DAILY and looks up config. No more hardcoded -1 in A_SOURCE_FILE_RECEIVED.' || CHR(10) ||
'v2.7.0 (2026-02-10): Added pRegisterExport parameter to EXPORT_TABLE_DATA_TO_CSV_BY_DATE. When TRUE, registers each exported CSV file in A_SOURCE_FILE_RECEIVED with metadata (CHECKSUM, CREATED, BYTES) from DBMS_CLOUD.LIST_OBJECTS. Enables file tracking and integrity verification.' || CHR(10) ||
'v2.6.3 (2026-01-28): COMPILATION FIX - Resolved ORA-00904 error in EXPORT_PARTITION_PARALLEL. SQLERRM and DBMS_UTILITY.FORMAT_ERROR_BACKTRACE cannot be used directly in SQL UPDATE statements. Now properly assigned to vgMsgTmp variable before UPDATE.' || CHR(10) ||
'v2.6.2 (2026-01-28): CRITICAL FIX - Race condition when multiple exports run simultaneously. Changed DELETE to filter by age (>24h) instead of deleting all COMPLETED chunks. Prevents concurrent sessions from deleting each other chunks. Session-safe cleanup with TASK_NAME filtering. Enables true parallel execution of multiple export jobs.' || CHR(10) ||
'v2.6.0 (2026-01-28): CRITICAL FIX - Added STATUS tracking to A_PARALLEL_EXPORT_CHUNKS table to prevent data duplication on retry. System now restarts ONLY failed partitions instead of re-exporting all data. Added ERROR_MESSAGE and EXPORT_TIMESTAMP columns for better error handling and monitoring. Prevents duplicate file creation when parallel tasks fail (e.g., 22 partitions with 16 threads, 3 failures no longer duplicates 19 successful exports).' || CHR(10) ||
'v2.5.0 (2026-01-26): Added recorddelimiter parameter with CRLF (CHR(13)||CHR(10)) for CSV exports to ensure Windows-compatible line endings. Improves cross-platform compatibility when CSV files are opened in Windows applications (Notepad, Excel).' || CHR(10) ||
'v2.4.0 (2026-01-11): Added pTemplateTableName parameter for per-column date format configuration. Implements dynamic query building with TO_CHAR for each date/timestamp column using FILE_MANAGER.GET_DATE_FORMAT. Supports 3-tier hierarchy: column-specific, template DEFAULT, global fallback. Eliminates single dateformat limitation of DBMS_CLOUD.EXPORT_DATA.' || CHR(10) ||
'v2.3.0 (2025-12-20): Added parallel partition processing using DBMS_PARALLEL_EXECUTE. New pParallelDegree parameter (1-16, default 1) for EXPORT_TABLE_DATA_BY_DATE and EXPORT_TABLE_DATA_TO_CSV_BY_DATE procedures. Each year/month partition processed in separate thread for improved performance.' || CHR(10) ||
'v2.2.0 (2025-12-19): DRY refactoring - extracted shared helper functions (sanitizeFilename, VALIDATE_TABLE_AND_COLUMNS, GET_PARTITIONS, EXPORT_SINGLE_PARTITION worker procedure). Reduced code duplication by ~400 lines. Prepared architecture for v2.3.0 parallel processing.' || CHR(10) ||
'v2.1.1 (2025-12-04): Fixed JOIN column reference A_WORKFLOW_HISTORY_KEY -> A_ETL_LOAD_SET_KEY, added consistent column mapping and dynamic column list to EXPORT_TABLE_DATA procedure, enhanced DEBUG logging for all export operations' || CHR(10) ||
'v2.1.0 (2025-10-22): Added version tracking and PARTITION_YEAR/PARTITION_MONTH support' || CHR(10) ||
'v2.0.0 (2025-10-01): Separated export functionality from FILE_MANAGER package' || CHR(10);
cgBL CONSTANT VARCHAR2(2) := CHR(13)||CHR(10);
vgMsgTmp VARCHAR2(32000);
---------------------------------------------------------------------------------------------------------------------------
@@ -149,8 +148,7 @@ AS
* Supports parallel partition processing via pParallelDegree parameter (1-16).
* File naming pattern: {pFileName}_YYYYMM.csv or {TABLENAME}_YYYYMM.csv (if pFileName is NULL)
* When pRegisterExport=TRUE, successfully exported files are registered in:
* - CT_MRDS.A_WORKFLOW_HISTORY (one record per YEAR/MONTH with export timestamp)
* - CT_MRDS.A_SOURCE_FILE_RECEIVED (tracks file location and partition info)
* - CT_MRDS.A_SOURCE_FILE_RECEIVED (tracks file location, size, checksum, and metadata)
* @example
* begin
* -- With custom filename
@@ -164,7 +162,7 @@ AS
* pMinDate => DATE '2024-01-01',
* pMaxDate => SYSDATE,
* pParallelDegree => 8, -- Optional, default 1, range 1-16
* pRegisterExport => TRUE -- Optional, default FALSE, registers to A_WORKFLOW_HISTORY and A_SOURCE_FILE_RECEIVED
* pRegisterExport => TRUE -- Optional, default FALSE, registers to A_SOURCE_FILE_RECEIVED
* );
*
* -- With auto-generated filename (based on table name only)
@@ -175,8 +173,8 @@ AS
* pBucketArea => 'ARCHIVE',
* pFolderName => 'exports',
* pMinDate => DATE '2025-09-01',
* pMaxDate => DATE '2025-09-17',
* pRegisterExport => TRUE -- Registers each export to tracking tables
* pMaxDate => DATE '2025-09-17'
* pRegisterExport => TRUE -- Registers each export to A_SOURCE_FILE_RECEIVED table
* );
* -- This will create files like: AGGREGATED_ALLOTMENT_202509.csv, etc.
* pBucketArea parameter accepts: 'INBOX', 'ODS', 'DATA', 'ARCHIVE'