Editing household IDs

[Zuhumnan Dapel]

Dear Users,
I have the following household IDs:

ID
0009f6b339824d57a39f5696c9c385a6
0011e7ca5a064a59a100b64c205d9689
0024ae5ebbdd417e95c31e18c90a1ab6
002e912ee9974a1987e27b1bebf909e3
0031d76399104961b5a58c71722f84f2
0035f38396e14cb99cae6f2fa3c14e96
0038ab99d05a433e946d077f166cfb89
003f5cb4aaa04e5bad13ce6b8b20357e
0043b1c8d9eb4bb78ab563f9fc7bf1b9
004d58321bda4aa68fa1249f1089666c

Any code on how to remove the numeric characters?

Thank you,
Dapel

[Zuhumnan Dapel]

I was able to get around this.

First, identify the set up of the ID using:

charlist ID
0123456789abcdef

This gives:

0123456789abcdef

Now is clear that numeric characters from

a to f

.

Finally, run

destring, generate(ID2) ignore(abcdef)

And boom

ID: characters a b c d e f removed; ID2 generated as double

[Joseph Coveney]

I'm not sure that it was good to do that. They look like concatenated hexadecimal numbers. IDs are generally okay (even best) as strings. If you need a corresponding numeric variable, for example, for use in some hierarchical regression command, then you could use -encode- for that.

[Zuhumnan Dapel]

Thanks. I'm not sure the IDs were originally so. I've been under the impression that the IDs were corrupted in the process of converting the file from SPSS to Stata.

Dapel

[Zuhumnan Dapel]

Possible to merge files using a string variable in both files?

[Joseph Coveney]

Yes, of course.

[Zuhumnan Dapel]

Ok. Why then this error message:

variable parentid1 does not uniquely identify observations in the using data
r(459);

?

[Joseph Coveney]

Because you have duplicate values for parentid1 in your using dataset. The error message has nothing to do with the datatype.

[Zuhumnan Dapel]

Thanks for flagging this. Any code for identifying and dropping the duplicates?

[Red Owl]

You can identify duplicates using this example:

clear
input str3 id byte var1
  001 17
  002 12
  002 14
  002 03
  003 10
  004 16
end


sort id

list if id == id[_n+1] | id == id[_n-1]

Then you will need to decide which observations should be dropped and can use -drop- as normal to drop unwanted observations.

Red Owl
Stata/IC 16.0 (Windows 10, 64-bit)

[Joseph Coveney]

Also, there's this:

help duplicates

[Nick Cox]

Let’s back up here. You don’t like your identifiers for some reason. Say you want shorter, simpler identifiers. So you remove some characters. But now there are duplicates. That shows that the removal of characters messed up your identifiers: they no longer are distinct. The solution is not to remove duplicates but to use a different method.

For example if you had Dapel1 and Dapel2 and then found yourself with two instances of Dapel, removing one won’t help.

That was the original question. It seems that you wanted to remove non-numeric characters, but the same point arises.

You’d do better to map your identifiers to dIstinct integers 1 up using egen. See the longstanding Stata FAQ about identifiers.

i am guessing here as there is no explanation of why you want to do this, but the thread is becoming an instance of the x-y problem; you are asking about y but what was the original problem x?