跳到內容

Text to Table Tips for Better Results

本頁內容尚未翻譯。

Use these tips when you want cleaner first-pass results and less manual cleanup afterward.

Text to Table uses the first parsed row as headers. It does not invent column names automatically.

If you want specific column names, include them in the first row before you convert the text.

This is especially useful for:

  • copied notes
  • contact lists
  • email summaries
  • AI-generated text

Structured text works best when the same separator is used consistently across the selection.

Good examples:

  • commas for CSV-style text
  • tabs for pasted spreadsheet rows
  • semicolons or pipes for repeated records

If the text mixes several separator styles, use Open in Table Editor so you can review the result before you apply it.

If one value contains commas, wrap that value in double quotes so it stays in one cell instead of splitting across columns.

Example:

For Markdown tables, keep the separator line between the header row and the body rows.

This is the reliable pattern:

  • header row
  • separator row with dashes
  • one row per record

If you copy Markdown from documentation or AI output, review the separator line before converting.

Choose Open in Table Editor when:

  • the source text is messy or inconsistent
  • you want to check headers before applying
  • you want to format the table before it is inserted
  • you are importing a file and want to confirm the result first

Start From File When the Data Already Exists

Section titled “Start From File When the Data Already Exists”

If the source is already saved as a file, start without selecting text and open Table Editor first.

Supported file types:

  • .csv
  • .tsv
  • .txt
  • .xlsx

File imports support local files and Google Drive™ files, with a 5 MB limit per file.

Use Image to Table for Image-Based Sources

Section titled “Use Image to Table for Image-Based Sources”

If your source is a screenshot, photo, or scanned page, use Image to Table instead of trying to run Text to Table on copied image text.

That path is better suited for OCR-based extraction.