How to identify unused fields

[StevenLawlor]

Hi,

I have inherited a workflow which pulls data from a very wide data source (200+ fields). Currently, most of these fields are brought the entire way through the workflow despite not being required as part of the output or not used at all during the workflow.

Is there a way to identify these so that I can remove fields as early in the workflow as possible?

My only thinking at the moment is using the Field Info tool and comparing but unfortunately there are also a lot of renames (multiple times) within the workflow also.

Thanks

[markcurry]

Hi @StevenLawlor 

If you've an newer version of Alteryx (2020.1 I think), the Data Cleansing tool has been updated to include an option to Remove null columns (and null rows).  

Or if you've an older version, you can use @MarqueeCrew 's CReW Delete Empty Fields macro tool available here: 

https://gallery.alteryx.com/#!app/CReW-Delete-Empty-Fields/5ab9138a0462d71894a25060

Hope that helps,

Mark

[StevenLawlor]

Hi Mark,

Thanks for the response. I should have been more clear, all of the fields have values however are not required as part of the output. What I want to do is figure out which fields I can remove at the start of the workflow because they are not in the output but also not required for any formulas etc. It's a large workflow with lots of formulas so manually going through each to note down every field which is used would be quite slow.

Hopefully I am explaining myself properly.

Steven

[markcurry]

Gotcha, sounds like a tricky one so.  Hopefully someone has a workflow that reads the workflow XML, and can identify the fields that tools like the formula tool use during the workflow.

[EnglishmaninNY]

how do I upvote this more?

I'm thinking there should be a button on certain tools (those with a list of the field contained). The button will cause an analysis of dependencies. I'm sure it will get more complex as you dig into it but for my typical use case I'm thinking two columns with a count which you can click to see detail:

- Column 1 - Count of use in prior join tool(s)

- Column 2 - Count of use in prior formula tool(s)

If it does fit in either of these buckets it feels like its just data which is flowing through but will not impact the selection of records you end up with.

What I find really frustrating is that if you input changes then any new fields just litter the whole workflow and without know which will have a material impact you have to track through all the tools to figure out where/if it is used. Or is there a trick to make this easier?

[aoxomoxoa]

i have a need for this as well -- has it been taken into consideration ?

to identify fields not used within a workflow at all