Streaming xls files Using Auto Loader

[YS1]

Hello,

Is there a way to read .xls files using auto loader or is there any workaround since excel files are not supported by the auto loader per the following document?

https://docs.databricks.com/en/ingestion/auto-loader/options.html

Thanks.

[shan_chandra]

@YS1 - Can you please check if you can convert .xls to .csv files and ingest using autoloader?

[YS1]

@Isha I tried to read them directly as csv files using different delimiters but the files weren't parsed correctly. The problem is that you have to check for the files that weren't read by the auto loader and then convert them to csv which is not the target here of the auto loader. My main goal is to read the files as they arrive to my folder without the need to check which one was read before.

[jose_gonzalez]

You can try to read the file as text file, using "wholeText" option https://docs.databricks.com/en/ingestion/auto-loader/options.html#text-options and then the whole file contents will be passed to the from_xml function and it will return an array of the records. Give it a try.

[YS1]

@jose_gonzalez I tried that, but it's giving me weird values. not sure what's wrong with that. 

[Sicnarf]

I am facing the same issue--I have a stream that I'd like to use autoloader on with an .xlsx. Is there any update to any workarounds on this issue?

[YS1]

I couldn't find any. I had to change the type of the files.