[langchain-sqlserver] Error when inserting a lot of documents

[yorek]

When adding a lot of documents to the vectorstore using the from_documents function, I get the following error:

2024-12-02 11:02:35,662 - ERROR - Add text failed:
 ('07002', '[07002] [Microsoft][ODBC Driver 18 for SQL Server]COUNT field incorrect or syntax error (0) (SQLExecDirectW)')

Traceback (most recent call last):
  File "W:\_git\_owned\azure-sql-langchain\samples\agentic-rag.py", line 47, in <module>
    vector_store = SQLServer_VectorStore.from_documents(
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "W:\_git\_owned\azure-sql-langchain\.venv\Lib\site-packages\langchain_sqlserver\vectorstores.py", line 578, in from_documents
    store.add_texts(texts, metadatas, ids, **kwargs)
  File "W:\_git\_owned\azure-sql-langchain\.venv\Lib\site-packages\langchain_sqlserver\vectorstores.py", line 836, in add_texts
    return self._insert_embeddings(texts, embedded_texts, metadatas, ids)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "W:\_git\_owned\azure-sql-langchain\.venv\Lib\site-packages\langchain_sqlserver\vectorstores.py", line 1239, in _insert_embeddings     
    raise Exception(e.__cause__) from None
Exception: ('07002', '[07002] [Microsoft][ODBC Driver 18 for SQL Server]COUNT field incorrect or syntax error (0) (SQLExecDirectW)')

It seems that if I try to insert less than 500 documents it works fine. With more it just return the above error.

[anushakolan]

From the offline discussion, the plan is to use batching to insert the documents into the database.

1. The batch_size an optional parameter for all the functions that allows inserting documents / data.
2. The default batch_size will be 100.

[anushakolan]

526 is giving this error message.

ERROR root:vectorstores.py:1308 Add text failed: ('07002', '[07002] [Microsoft][ODBC Driver 17 for SQL Server]COUNT field incorrect or syntax error (0) (SQLExecDirectW)')

[anushakolan]

But inserting 525 documents is giving a different error message,

ERROR root:vectorstores.py:1308 Add text failed: ('42000', '[42000] [Microsoft][ODBC Driver 17 for SQL Server][SQL Server]The incoming request has too many parameters. The server supports a maximum of 2100 parameters. Reduce the number of parameters and resend the request. (8003) (SQLExecDirectW)')

[anushakolan]

Now inserting 524 documents is working fine,

[anushakolan]

the number of parameters are calculated as 5 * (No_of_documents) + 1, as shown in above image, we get the following parameters for every document,

id: None custom_id_m0: 1 content_metadata_m0: {'id': 1, 'title': 'Document 1'} content_m0: This is the content of the page. embeddingvalues_1: [1.100672896551693, 0.6787781106600161, 0.2603396374499168]

And one additional parameter for embedding_length.

So for 5 documents in above case, we get 5 * 5 + 1, 26 parameters.