converting string to numerical

[Ellen Kiley]

I just uploaded someone's Excel datafile, and there are a bunch of categorical variables which are numbered in Excel but show up in Stata as string. When I "destring" the variables this is the error message I get:
Age: contains nonnumeric characters; no replace
Race: contains nonnumeric characters; no replace
Ethnicity: contains nonnumeric characters; no replace
RaceEthnicityCoding: contains nonnumeric characters; no replace
PatientStatus: contains nonnumeric characters; no replace
Barrier: contains nonnumeric characters; no replace

Is it because the Excel cells include blank spaces? If so, how do I fix this? All ideas appreciated thank you!

[Nick Cox]

We can't see any of your data to tell you what the non-numeric characters are. See e.g. #5 of https://www.statalist.org/forums/for...iable-problems for a command you can use, but the main idea is simpler:

foreach v in Age Race Ethnicity RaceEthnicityCoding PatientStatus Barrier {
     di "`v'"
     di
     tab `v' if missing(strofreal(`v'))
}

That is, you need Stata to show you, and you can then show us, which values of each variable could not converted simply to numeric values.

The answer is likely to be to use destring with particular options.

[Ellen Kiley]

Thank you. Solved

[Nick Cox]

Excellent, but a hint at what the problem was and how you solved it might be of use or interest to people finding this thread in future.

[Ellen Kiley]

I am using someone else's dataset and rushing and she had instructed me to import the whole sheet without mentioning that way at the bottom out of obvious view she had done calculations/etc in Excel which screwed up data! Sorry for the false alarm.

[Ellen Kiley]

Or I can just use Pearson's (which runs without error) because while 2 cells have only 5 observations, only 1 of the expected has <5 and that is way < 20%?

[Nick Cox]

#6 is posted in the wrong thread.

[Ellen Kiley]

Anyhow thanks for your help I really appreciate it