When Table 1 spans across multiple pages, the current parsing pipeline may lose structural continuity, resulting in incorrect or incomplete extraction.
Observed behavior
- Table content is split at page boundaries
- Rows and columns become misaligned after the page break
- Header or column definitions may not be consistently propagated
- In some cases, the second half of the table is treated as a separate or unrelated structure
Expected behavior
- Multi-page tables should be detected as a single logical table
- Structural consistency (headers, column alignment) should be preserved across pages
- Table reconstruction should maintain row continuity and semantic coherence
When Table 1 spans across multiple pages, the current parsing pipeline may lose structural continuity, resulting in incorrect or incomplete extraction.
Observed behavior
Expected behavior