Marrying in-coming CSV data with metadata from a sister file in the same series

Question

Hello--

Thanks so much for taking my question.

Occasionally I have to read in a large, thousand-column file in with CSV format.   Is it possible instead of using Auto-Field to attempt large-scale metadata assignments that fail frequently, to instead "lend" the metadata from a sister file in the same series?   I see literally hundreds of fails using Auto-Field, and would really appreciate a better way to assign metadata values in a more consistent manner.

Thanks again,

Michael M.

manion72 · Answer

Hi @ArtApa, thanks for your suggestion last month, I appreciate it...  Just wanted you to know I did read it.  Unfortunately my file is actually a tab-delimited, but I learned something about flat-files, so appreciate your post!

ArtApa · Answer

Hi @manion72  - You can also download metadata in the Input Data tool:

In order to do that, you would need to read your csv as flat and then download Flat File Layout: https://help.alteryx.com/current/designer/flat-file-layout

manion72 · Answer

Hi Joshua--

Thanks for the suggestion, yes that's definitely a possibility.   My issue is I actually have two of these wide series of files and another with about 250 fields, and not a lot of time since my company just decom'ed SAS which I've used for 7 years, leaving me to pick up the pieces of my processes in Alteryx...!  😉

But if you hear of any other tricks or hacks, I'd still be interested.  For example, one need I had a few weeks ago was to take Column Order from one file and use it to order a second in the same series, which did indeed work as long as the literal connection to the Union Tool was first made to the file I was using to source the field order.   I tried that trick for metadata too, but unfortunately it didn't work:  I unioned my CSV with a yxdb dataset, making the yxdb #1 in ordering, and connecting the yxdb to the union tool first.  The result each time though was to default all fields from the resulting union to the default string type.   Maybe it takes the lowest common denominator for each field, not sure.

Appreciate your suggestion, I will keep it in mind.   It would just be great to be able to leverage a "help" to avoid a mandatory front-load effort whenever I have a CSV from an existing process, lol.

Happy Holidays!  🙂

Michael M.