Map a table

IQ Bot uses column names as reference points when extracting values from tables. As part of training a learning instance in the Designer, you must map the column names and configure the table settings.

Prerequisites

Create a learning instance and map all the form fields in the Designer. For more details, see Map a field.

Procedure

  1. Skipping the Table settings section, click the first table column in the Table fields section.
  2. Click the name of the column from where you want to extract data in the image display panel.
    There are two ways to map the column values:
    • Option 1: For a basic table that is contained within one page, map the column name to the Column value field.
    • Option 2: If the table spans multiple pages, we recommend mapping the Column value field to an entry in that column that has the greatest height and width among the entries in that column.
    Depending on which mapping option you choose, IQ Bot autopopulates the Column value field in the field details panel with either the name of the selected column or the value of the selected table entry. In addition, the selected SIR in the document is highlighted with yellow color for easy identification.
  3. Optional: Drag the corners of the green-bounded box to resize the mapping area.
    Alternatively, draw a new mapping area by holding and dragging the cursor.
    When you resize any value region or create your own value region using select, its position is fixed relative to the field label. The value for that field is always searched in that relative region.
    Reset a fixed-field region defined by the user in the autodetected mode by clicking the close button at the top-right corner of the field.
    Tip: For additional information, see How to map tables in IQ Bot (A-People login required)
  4. Expand the Field options section and select Required or Optional. Use the optional value if the column value is not consistently available in the training documents.
  5. Optional: Provide validation patterns.
    Designer validation patterns

Configure the Table settings section. Ensure that you map all the other fields in the document before populating this section.

  1. Review the selected option in the Best field for table/repeated section field. IQ Bot auto detects the table field that best defines the height of the rows in the table. You can select a different column from the list.
    Other columns are extracted in reference to this column. For example, if the Best field for table/repeated section has five rows, then a maximum of five rows will be extracted from the other columns.
    Note:
    • Ensure that this field does not contain multiple lines or is empty.
    • Sometimes, the system suggestions are not updated when you validate the documents.

      Workaround: You must update the correction manually in the validator for IQ Bot Cloud if the system suggestions are not displayed.

  2. Optional: Map a field label for the End of table indicator field to specify where to stop extracting data from that table.
    For example, an invoice might contain the subtotal, sales tax, and total in the table, but because each of these fields has only one value per document, they should not be extracted as part of the table.
    Provide the name of the first field that appears where the table ends, for example, Total or Grand Total.
    Provide the value for the End of table/section indicator field in any of the following ways:
    • In the Image display panel, select an SIR to populate that text value as the End of table/section indicator.
    • Alternatively, type a value directly into the End of table indicator/section field, in the middle panel.
    Tip: Use the pipe symbol (|) to add field aliases, which represent alternative text that might be in an SIR. For example, Subtotal | Subtotal Amounts refers to Subtotal OR Subtotal Amounts.
    For an example of how to map one table for extraction in a document that contains multiple tables, see Mapping a table in a page with many tables.
  3. If you mapped a table value for the Column value field (option 2 in step 2): Expand the Advanced table options section and select Stop extraction at "end of table indicator".
  4. Optional: Enter custom logic.
    Adding custom logic to improve automatic extraction in production

Next steps

Click See extraction results to check the training results for table data and remap the fields if required. Preview extracted data