DB Query Node Reading Partially From Data Warehouse

Hi All,

I am trying to read tables from a data warehouse. The warehouse is SAP BI using Sybase. The node is able to see some columns within the tables but for a number of tables, I get the error “error fetching columns”. What may the problem be?

Thanks.

Still stuck here. Any guidance to resolve this is welcome.

Hi @Steve_Bunde , what form do the table and column names take? Is it every table, or just specific tables that have the problem? Anything special about their names? Do they contain any characters other than alphanumeric and underscore?

I haven’t got much to go on (I don’t use SAP BI/Sybase), but I’m wondering if your problem may be similar to this:

2 Likes

Hi @takbb

Thanks for the response. I am using the DB Connector node in this case. I have noted that it has the parameters mentioned in the topic you shared that references the Google BigQuery Connector. I
a) In the DB Connector, I went to the Advanced tab.
b) There, in the “Dialect syntax” section, I found the point “Replace non-word characters in identifiers”, which has a tick in the default setting.

However, I’m still getting the same “error fetching columns” message.

I’ve tried checking the naming of the tables but there seems to be no difference in what is read successfully and the ones that fail.

What may I still be missing?

We have not much to go on here. Do you use an official driver? Do you have all the access rights?

SELECT column_name FROM information_schema.columns WHERE table_name = 'YOUR_TABLE';

I notice that the table in question seems to have upper and lower cases and special characters in the table name.

2 Likes

Thanks @mlauber71 for your response.

Just to clarify, I have full access rights and use the official driver from Sybase (that’s why some tables are visible). I did check your suggestion on characters and cases which gave a different point of consideration. I have however no pattern as to why some can be read while some return an error. Let me pursue the angle on naming as suggested by both yourself and @takbb

1 Like