Fix for Parquet Import Failure Due to Column Count Mismatch Across Files
Details
| Detail name | Value |
|---|---|
| Changelog Number | 29559 |
| Type | Bug |
| Status | Resolved |
| Affected Versions | Exasol 2025.1.0 |
| Fix Versions | Exasol 2026.1.0, Exasol 2025.2.1, Exasol 2025.1.11 |
| Resolution Date | 2026-03-20 |
Description
When performing an IMPORT INTO operation with multiple Parquet files, the import operation fails if one or more of the subsequent files contain a different number of columns compared to the first file's column count. This issue arises due to schema mismatches between files during batch imports, causing the process to crash without a descriptive error message to indicate the root cause.
Workaround
- Option 1: Group files with similar schema together and execute the IMPORT INTO operation for each schema group separately. This avoids schema mismatch issues during batch imports.
- Option 2: Use the Parquet column mapping feature to map columns from each file explicitly to ensure consistent schema before importing.
Fix
The fix introduces additional validation logic to detect schema mismatches between files during the IMPORT INTO operation. If a mismatch in column counts or types is detected, the process will terminate with a clear and descriptive error message, indicating the root cause of the failure. Furthermore, the process now includes safeguards to prevent crashes or undefined behavior resulting from inconsistent schemas in input files.
