Reliable Text Extraction for DOCX, XLSX and PPTX Files in Document Indexing

· Web Crawler · Bug Fix · All updates

A document that indexes with empty text is worse than one that fails outright: it looks fine in the index and never matches anything. Document indexing now extracts text reliably from all the major office formats.

01 What was going wrong

Two causes, both silent. Some files had their format misdetected and were parsed as something they were not. Others hit encoding issues that produced nothing usable. In both cases the document reached the index with empty text.

Before

Many office documents were indexed with empty text, because of format misdetection and encoding issues.

After

DOCX, DOC, XLSX, XLS and PPTX are fully supported, with the text extracted properly.

02 What gets extracted

Extraction is not limited to the main body. Tables, headers and footers are included, as are speaker notes in presentations — the places where the answer to a search often actually lives.

FormatStatus
DOCX, DOCFully supported, including tables, headers and footers.
XLSX, XLSFully supported.
PPTXFully supported, including speaker notes.
View the full changelog