Row Match

Can be one of the following types:

The 'Row Extractor' defines how each table row is identified and how cell values are extracted. This extractor should return a 2D result, where each hit contains named children—one for each Data Column to be populated.

• Use a Pattern Match extractor with named groups to map extracted values to columns. For example: (?<Quantity>\d+) (?<Unit_Price>[\d,.]+) (?<Total>[\d,.]+) This will capture cell values for the "Quantity", "Unit Price", and "Total" columns. Note: spaces are not allowed in regex group names; use underscores (_) instead.
• Alternatively, use a Data Type with child extractors named after the columns, and apply 2D collation methods such as Array, Ordered Array, or Pattern-Based to cluster hits into rows.
• If a column's value is not present in the extractor result, and the column has a Value Extractor defined, it will be executed against the full row content to populate the cell.

Proper configuration of the 'Row Extractor' is essential for accurate table extraction, especially in documents with irregular or headerless tables.

Can be one of the following types:

The 'Header Extractor' defines the start of the table content. Any matches occurring prior to the header will be discarded. When set, this property also enables the use of 'Maximum Header Spacing' to limit the distance between the header and the first row.

• If using Label Sets, specify the header label for the table instead of setting this property, and ensure the 'Use Labelset' option is enabled.
• Use a Pattern Match or other extractor to target the specific text or pattern that marks the table's header.

Configuring the header extractor helps prevent accidental inclusion of non-table content and improves extraction reliability in documents with multiple tables or complex layouts.

Can be one of the following types:

The 'Footer Extractor' defines the end of the table content. Any matches occurring after the footer will be discarded.

• Use a Pattern Match or other extractor to identify the text or pattern that marks the table's footer.
• If using Label Sets, specify the footer label for the table instead of setting this property, and enable the 'Use Labelset' option.

Setting a footer extractor is useful for tables with summary rows, totals, or other content that should not be included as data rows.

The 'Maximum Header Spacing' property limits the allowed vertical gap between the detected header and the first data row. This value is expressed as a percentage of the average font size.

• A value of 0% disables this check (no maximum).
• If set, the spacing between the header and the first row must be less than or equal to this value for the table to be considered valid.

Use this property to prevent accidental inclusion of distant or unrelated content as table rows, especially in documents with variable spacing or multiple tables.

The 'Maximum Row Spacing' property controls the allowed vertical gap between consecutive table rows, as a percentage of the average font size.

• A value of 0% disables this check (no maximum).
• If set, the table will be terminated if the distance to the next row exceeds this value.

Adjust this property to handle tables with variable row spacing, or to prevent inclusion of unrelated content when tables are separated by large gaps.

Can be one of the following types:

A combination of the following flags:

The RowMatchOptions enumeration provides flags that enable or disable specific features of the Row Match extraction method. By combining these options, you can fine-tune how tables are detected, how rows are processed, and how control characters are handled.

Use these options to adapt Row Match extraction to the structure and requirements of your documents. For example, you can prevent tables from spanning multiple pages, enable recursive extraction for empty columns, or use label sets for header/footer detection.

Consider the impact of each option on extraction results and performance, and adjust as needed for your specific use case.