Hi All,
I am having trouble aggregating data as I constantly have to mess with the string vs data formats.
I am having issues where I have info that should be very clearly a data but is being read by Knime as a string.
I know I can use nodes to convert this, but would be easier if knime read it correctly the first time. Any Tips?
Thanks,
Hi @rparr009 , can you clarify what you refer to as “data”?
Data can be stored as string, so I’m not understanding what you mean by “I have info that should be very clearly a data but is being read by Knime as a string.”
By the way, based on the data you presented in your screen shot @rparr009 , these columns are coming out as string because of the missing values (identified as “?”), which are basically NULLs.
@rparr009 you might want to check where the data is being imported. It might be that the decimal separator is not correctly interpreted so every number with a “.” is seen as a string. If it is the CSV reader you can adjust the settings.
In general it does make sens to store data in mor ‘stable’ formats than CSV (ARFF, Parquet or ORC could be alternatives, also H2 and SQLite - all of which are support ed by KNIME)
When I say data, I mean numbers.
I am currently using and excel reader for this. How would I adjust the settings?
Thanks for the clarification @rparr009 . As I mentioned, this is probably happening because of the missing values. When the file is generated, can the missing values be replaced by 0 instead of just empty cells?
Otherwise you will have to do this replace in Knime, and then change the column types, which you are trying to avoid.
Have you tried changing the type with only one node for all columns. Then the additional work is rather limited.
Hi @rparr009,
is it possible to upload a sample of your data and workflow to see how you read your data (excel reader settings)?
This make it easier to identify the reason for your problem.
BR
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.
