Text Mining action fails due to fields sized too small

SYMPTOMS

In the settings of the legalreview_v31 template, certain dictionaries are included. If the size of the fields that are mentioned are too small, and if a single document of a set you are processing is generating a field entry that is bigger than this dictionary field size, the entire set that is indexed at that time will fail. The tasks end up completed and green in the processing interface. No records  are created in the legalreview database tables.

CAUSE

The tasks end up completed and green in the processing interface. No records however are created in the legalreview database tables.

RESOLUTION

To solve this problem the field size has to changed to allow more characters to be stored.

The DictionaryInfoDto section of the Text Mining index can changed from 20 to 4000 <DictionaryInfoDto> <Name>PTMURIIP</Name> <ValueField> <DbType>NVarChar</DbType> <Size>20</Size> <IsIndexed>true</IsIndexed> </ValueField> </DictionaryInfoDto>

This will allow a larger amount of entity matches to be stored.

APPLIES TO

3.1