Appendix H: Limitations of the Gateway

Limitation 1:

The Gateway does not support the vertical text extraction or column-wise text extraction by default extraction (that is, IFilter apply="false" or is undefined in the Extract configuration XML file).

The extraction of texts is done from top-left to bottom-right resulting in the text objects, as shown below:

A total of nine text objects are extracted. Text on the vertical axis is not recognized as a single object as the extraction is done horizontally.

Note: The text can be extracted as a single object when using iFilter.

Limitation 2

Default processing of PDF files (that is, IFilter apply="false" or undefined in the Extract configuration XML file) results in objects with Page attributes. There is a limitation where the Page index starts with 0 irrespective of the page numbers displayed in the PDF Document.

Limitation 3

The Gateway is unable to detect whether the text in the input PDF is encoded and can only extract the text 'as is'. You can determine if the text is encoded by opening the file in a PDF reader, selecting some of it and copying it to a plain text application such as Notepad. If the plain form of the text is a set of scrambled symbols, then the input document's text is encoded.

Limitation 4 - Extracting Large Strings of Excel Files Using ODBC

Text can be cropped when using ODBC to extract large strings from Excel files.

Cells in Excel can have more than 255 characters, but ODBC expects such large cells based only on cells in the first N rows, defined by TypeGuessRows (default = 8) in the Registry. This is one of the Access Connectivity Engine parameters located in the Regsitry at: Computer\HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Office\ClickToRun\REGISTRY\MACHINE\Software\Microsoft\Office\16.0\Access Connectivity Engine\Engines\Excel.

Therefore, if TypeGuessRows is set to 8 and the first large cell is in a row after row 8, then the extraction from ODBC will only read the first 255 characters.

You can use one of the two workarounds:

• Use a Datasource of 'Excel' rather than the ODBC driver, this requires MS Office to be installed.

• Set TypeGuessRows in the Registry Editor to 0, but this may increase the processing time and/or memory consumption.

Limitation 5 - Supporting Non-English User Locales

When the Gateway is installed on a system where the user has selected a non-English locale that uses a comma as the decimal separator (that is, a number 123.45 in a Windows system with a French locale would be expressed as 123,45), the Gateway does not interpret such numbering formats from input sources and configurations nor does it export them into output files.