Database search

Hello,

I have a little problem with the pubchem database. In the first part of my final workflow I am extracting CID based on the exact mass and the second querying is extracting the molecular formula based on CID.
Sometimes (depending on the time of day - it seems like) I end up in the node “get request” column “body” with few rows empty which has an impact on the performance of my workflow. The second extraction of the molecular formula starts with a chunk loop and parallel chunk loop following the get request node, json path and json to table. However, Parallel chunk loop fails (all the time) with the message; “Not all chunks finished. - check individual chunk branches for details”. When I open metanode, “json path” node fails with the message that there is no json column.

I have two questions: How can I avoid missing information in the “body” column and the second question, how can I avoid failed execution?

Thank you in advance

Hi,

Welcome to KNIME Forum. You can use Missing Value node to handle the empty rows, you can either set the node to treat strings or number in a certain way as indicated in the screenshot if you have a dynamic data stream:

You may also treat individual columns if you have fixed column specs:

In order to avoid failed execution, you may use the Try and Catch Errors nodes in the pipeline. Here is an example from the KNIME HUB to get you started.

Best regards,
Hannan

1 Like