Dynamically update Datatypes

Hi,

Is there a way when new columns come into my workflow, they can automatically update to be a fixed decimal or any type of number field? They are currently coming in as V_String and I was hoping that this would be something I wouldn't have to do manually.

Dynamic Processing

Accepted answers

All comments

CoG

The Auto Field Tool may be useful for you.

kamels

Do you have any advice on how to use that? I could not figure out how to use that tool to create a Fixed Decimal, when I ran the workflow nothing changed.

jrlindem

Like @CoG recommended, the Auto Field Tool is a good way to accomplish this. That said, if there are characters in your field that don't cause Alteryx to automatically detect it as a number field, it's not going to assign Fixed Decimal or Double or Int types....

Another alternative is to use the Field Info tool to help identify the fields that are coming into your workflow, then you can flag them using a formula tool and convert them using the Multi-Field Formula Tool, using a formula like ToNumber([field]) to get the data to act as numbers. This is a lot more complex, though and harder to illustrate. Recommend doing some searching on this combo of tools to get what you want.

Hope that points you in the right direction. -Jay

flying008

Hi, @kamels

Could you please tell me, Was the name of new field is fix or dynamic ?

kamels

The field is dynamic. It is the month number and year combined. (For example, March 2025- "0325") It has all 12 months in it at the same time, but it is dynamic because the following year, all the dates will update to "0326" ect.

Diederik_vanderharst

@kamels
Put a text tool with dummy values for the known fields on your canvas and set the required data types via select tool. Then put a select records on it, and let no records pass on. After the select records place a union and union it with your regular input. If the stream from your "template" is first to connect to the union unioned streams will inherit that schema.
About the dummy values. If you know that they are months of the year, then you can prep 60 fields for the next 5 years, plus the standard fields, and have it union on common subset.

Good luck! Let me know how you did. See pics below. And don't forget, connect the template to the union first. Otherwise delete the union and try again.

Schermafbeelding 2025-10-09 om 14.00.19.png

Schermafbeelding 2025-10-09 om 13.56.39.png

Schermafbeelding 2025-10-09 om 14.02.47.png

Here's a quick example how to create that template. You can also make that more dynamic using a generate rows tool in stead of text input.

Schermafbeelding 2025-10-09 om 14.08.25.png
Last edit: I transform everything to V_string here, not noticing that you needed a number type. Should not matter obviously.

Not a fan of autofield. If data typing is important, do it yourself.

flying008

Hi, @kamels

FYI.

REGEX_Match([Name], '[01]\d2\d')

录制_2025_10_10_11_44_45_889.gif

apathetichell

Auto-field will not swap vstring to fixed_decimal ---> it will swap v_string to string with a size of the max_entries in the field.

CoG

@apathetichell

That is incorrect. If a string field has only numeric values, the Auto-field will convert it to a numerical type. If whole numbers, then one of the int types, otherwise a double. The OP mentioned Fixed Decimal or "any numerical type". This is precisely what the Auto-field tool was built to do. That being said, you are certainly correct that Fixed Decimal would require a more complex solution as others have already suggested.

apathetichell

@CoG --- you are correct. I was incorrect in the earlier statement. thanks for pointing that out. I would posit though that this is not what the auto field was tool was built to do --- and it was built to dump excess space in string fields because the difference between a vw_string default size and the max size in the field can be huge and this really hits performance in large datasets.

dreldrel

You could consider using multiple field to pre-process your data then use auto field to update the data type

kamels

This is the data that I am looking at. The Auto-Field will not work to correct this to a numeric value in order to create formulas.

Diederik_vanderharst

@kamels I'm not surprized because it looks like the top two records contain text.

The autofield will look at what is most efficient, but also what is possible. A field that contains as little as one text char will never become a numerical field. So either you clean up the data so you are left string fields that only contain numbers before transforming them to numerical. Or you have to calculate within string fields transforming every value to number before in row eg. tonumber([field1]) / tonumber([field2])

Quick Links

This months top contributors

atcodedog05 19598

Qiu 15878

binu_acs 15708

MarqueeCrew 13708

apathetichell 13703