Using the select tool to isolate leading number in a string

Hi,

I have a some strings with numbers at the start, e.g.

35.68]/dsfo48,.fdhgk

45joidfo8.980

And i'm trying to just extract the leading numbers, so the result for the above would look like;

35.68

I believe 1 way of doing this would be to use the RegEx tool, but it seems like I can achieve the same output by just changing the data type of the field from string to double using the select tool.

I was just wondering whether this is a reliable way of performing this operation? Or whether the Alteryx implementation of the data conversion could throw up issues in some cases?

Also, is there a repository where things like this are documented? I did try googling, but it seems there's only really documentation of how functions/tools operate on a higher level, not so much the "nuts and bolts" of the implementation.

Thanks

Best Practices

Accepted answers

BenMoss

It does appear to work, I wouldn't say this is a reliable solution, predominantly because of the rules that you would place on your incoming data.

What if the number you want isn't always the leading value in the string.

What if their are no letters separating the two numeric parts of the string.

If you are sure that you can apply the rule 'Alteryx will take only the first numbers of the string and drop everything else', then I guess it's fine.

Personally i'd go with regex even if I knew this was the case. You know it works. Have you actually measured the performance difference when using the two methods?

Ben

All comments

MarqueeCrew

@CB1,

Try using a Regular Expression like this:

Regex_Replace([String],"(\d+\.{0,1}\d+)\D.*",'$1')

Create a GROUP (indicated by the data enclosed in parenthesis)

That GROUP starts with 1 or more numbers (\d) and is immediately followed by a period (\.) that can occur 0 or 1 times ({0,1}) followed by 1 or more numbers.

This GROUP is followed by 0 or more Non-Numbers (\D*). That group is identified as GROUP #1.

Replace (or output) the data with the string found in GROUP #1 ($1).

Cheers,

Mark

CB1

Thanks for the reply Mark.

I appreciate that this can be done using RegEx, and I understand its implementation, but my question is more focused on whether the select tool is an acceptable and reliable solution.

Given that it's slightly simpler and quicker, my preference would be to use the select tool, which is why I'm asking.

Quick Links

This months top contributors

LanisC 335

hhaigh1 280

erich_shiino 280

FláviaB 262

JORGE4900 235