SplitSkill : Required skill input is missing or empty. Name: 'text', Source: '$(/document/content)'.

Question

SplitSkill : Required skill input is missing or empty. Name: 'text', Source: '$(/document/content)'.

Jacek 25

I'm using SharePoint as a Data source for my indexer.

I created SplitSkill to split the document/content for further AzureOpenAIEmbeddingSkill.

When Data Source is set to one library everything works great, but when indexing whole Share Point site, SplitSkill is crashing on one document:

"Could not execute skill because one or more skill input was invalid."

"Required skill input is missing or empty. Name: 'text', Source: '$(/document/content)'."

General indexer error is: A request to SharePoint timed out. Retrying indexer execution later may resolve the issue. To automatically resume execution consider adding a schedule to the indexer, see https://docs.microsoft.com/en-us/azure/search/search-howto-schedule-indexers

I did index the whole content with out enrichments, and worked fine, so I'm guessing there is no problem with timeouts. I did try to check which document is causing problems but when indexed could not find the one with out content. Indexer is returning the details about problematic file, but I can't decode it. I guessed that the second guid is id of the file, but when asked SharePoint about it, it returned nothing:

localId=siemens.sharepoint.com%2ccf581cde-9561-4c06-82aa-a96529bcc2d5%2c39bbc641-2f7e-409c-bc69-8ab7a28015d9%3a%3ab!3hxYz2GVBkyCqqllKbzC1UHGuzl-L5xAvGmKt6KAFdkOvVkQt22LRoFT4E01Zp_f%3a%3a01VCYAUHUEP3KY6OXDYFD3DDZH7KLYAFLY&documentKey=siemens.sharepoint.com%2ccf581cde-9561-4c06-82aa-a96529bcc2d5%2c39bbc641-2f7e-409c-bc69-8ab7a28015d9%3a%3ab!3hxYz2GVBkyCqqllKbzC1UHGuzl-L5xAvGmKt6KAFdkOvVkQt22LRoFT4E01Zp_f%3a%3a01VCYAUHUEP3KY6OXDYFD3DDZH7KLYAFLY

Ramr-msft 17,826 Reputation points

2024-01-03T10:54:12.5733333+00:00

Thanks for the question, Here are the troubleshooting steps that could help.

The SplitSkill Error: The error message “Could not execute skill because one or more skill input was invalid” and “Required skill input is missing or empty. Name: ‘text’, Source: ‘$(/document/content)’” indicates that the SplitSkill is not receiving the expected ‘text’ input from the document content. This could be due to a document that doesn’t have any content in the field that SplitSkill is trying to process. You might want to add a check in your pipeline to ensure that documents being processed by SplitSkill have the necessary ‘text’ content.
Jacek 25 Reputation points

2024-01-03T15:08:14.2833333+00:00
Well, not quite. I'd like to point out that the problematic documents are the presentations or PDF documents containing only pictures. The content of such a file consists only of multiple new lines ( '/n' ). It seems like SplitSkill cannot handle, for example, such content ->

"content": "\n\n",

Any idea how to deal with this issue?