Extract PDF Table

Extracts tabular data from a PDF using gopdf's table detection. Provide header anchors to locate the table precisely, or leave headers empty for auto-detection. Returns {columns, rows} where rows are header-keyed maps. Prefer this over RunPython with manual span geometry for any table-shaped PDF extraction (invoices, spec lists, statements). Handles multi-line cells (mergeGap), table-end detection (maxRowGap), and cross-page continuations.

Category: Structured Data Files (Excel, JSON, PDF) · Group: Files · Activity name: ExtractPDFTable

Example

outputVariable = ExtractPDFTable("<filePath>", headers=[], mergeGap=0.0, maxRowGap=0.0, anchorColumn="<anchorColumn>", requireAnyColumn=[])

Inputs

Name	Type	Required	Default	Description
PDF File Path	string	Yes		Path to the PDF file to extract the table from.
Header Anchors	array			Column header strings that identify the table's header row (case-insensitive substring match on the same Y-line). Leave empty for auto-detection via gap analysis.
Merge Gap (pt)	float			Maximum vertical distance between consecutive rows to merge them into a single logical row (multi-line cell support, e.g. bank statements). Default: 0 (no merging).
Max Row Gap (pt)	float			Maximum vertical distance between consecutive data rows; rows beyond this gap end the table (useful for stopping at footers). Default: 0 (no limit).
Anchor Column	string			Column name that signals the start of a new logical row. Consecutive rows where this column is empty merge into the previous row. Case-insensitive.
Require Any Column	array			Column names where at least one must be non-empty for a row to be included. Rows where all listed columns are empty are dropped.

Outputs

Name	Type	Description
outputVariable	json	JSON object with the detected table: - columns: Array of column names (in page order) - rows: Array of row objects keyed by column name Example: {{ map(table.rows, .["Suppliers Code"]) }} to pull one column; {{ filter(table.rows, .["Qty"] matches "[0-9]+") }} to filter.