Generate unique Identifier from 2 variables

[CAPM]

Hello,

I encountered the following problem:
Out of 2 existing variables( a date and a variable consisting of numbers and letters) stored as str6 (format: %9s) and long (format: %tdDDNNCCYY) I need to create an identifier by combining these 2 variables.

egen ident =concat(date var2)

yields this error code: factor variables and time-series variables not allowed.

Thanks a lot in advance!

[Joseph Coveney]

egen ident =concat(date var2)

yields this error code: factor variables and time-series variables not allowed.

No it doesn't.

.ÿversionÿ13.1

.ÿ
.ÿclearÿ*

.ÿsetÿmoreÿoff

.ÿ
.ÿinputÿstr6ÿvar2ÿlongÿdate

ÿÿÿÿÿÿÿÿÿÿvar2ÿÿÿÿÿÿÿÿÿÿdate
ÿÿ1.ÿ"abc123"ÿ19992
ÿÿ2.ÿ"DEF456"ÿ19993
ÿÿ3.ÿend

.ÿ
.ÿformatÿvar2ÿ%9s

.ÿformatÿdateÿ%tdDDNNCCYY

.ÿ
.ÿegenÿidentÿ=concat(dateÿvar2)ÿ

.ÿlist,ÿnoobs

ÿÿ+---------------------------------+
ÿÿ|ÿÿÿvar2ÿÿÿÿÿÿÿdateÿÿÿÿÿÿÿÿÿidentÿ|
ÿÿ|---------------------------------|
ÿÿ|ÿabc123ÿÿÿ26092014ÿÿÿ19992abc123ÿ|
ÿÿ|ÿDEF456ÿÿÿ27092014ÿÿÿ19993DEF456ÿ|
ÿÿ+---------------------------------+

.ÿ
.ÿexit

endÿofÿdo-file

Now, why don't you follow suit?

[Nick Cox]

I don't believe that! Your problem can't be reproduced:

 
. clear

. set obs 1
obs was 0, now 1

. gen long date = mdy(10,7,2014)

. gen var2 = "answer"

. egen ident = concat(date var2)

That works.

This post is reminiscent of http://www.stata-forum.de/stata-synt...ugen-t872.html where you used quite a different syntax, which was illegal, which was why you got an error message.

Please read FAQ Advice Section 5 on use of real names being preferred; Section 8 on cross-posting (you should tell us about it); and Section 12 on reporting exactly what you typed and exactly what happened.

[Joseph Coveney]

Your cross-post (I didn't know about it until Nick's post above) mentions matching, but maybe you mean merging? If so, then you can do that with the two variables as-is.

merge 1:1 var2 date using einem_anderen, keep(match)

Provided, of course, that the two variables exist separately in the other dataset.

[CAPM]

First of all thanks for your answer. Indeed I am a Stata beginner and am sry for the inconveniences. I will read the FAQs asap.

To your answer: What I need is this:

ÿ+---------------------------------+
ÿÿ|ÿÿÿvar2ÿÿÿÿÿÿÿdateÿÿÿÿÿÿÿÿÿidentÿ|
ÿÿ|---------------------------------|
ÿÿ|ÿabc123ÿÿÿ26092014ÿÿÿ26092014abc123ÿ|
ÿÿ|ÿDEF456ÿÿÿ27092014ÿÿÿ27092014DEF456ÿ|
ÿÿ+---------------------------------+

As can be seen in your example the date is displayed as a number instead:

ÿ+---------------------------------+
ÿÿ|ÿÿÿvar2ÿÿÿÿÿÿÿdateÿÿÿÿÿÿÿÿÿidentÿ|
ÿÿ|---------------------------------|
ÿÿ|ÿabc123ÿÿÿ26092014ÿÿÿ19992abc123ÿ|
ÿÿ|ÿDEF456ÿÿÿ27092014ÿÿÿ19993DEF456ÿ|
ÿÿ+---------------------------------+

Is there a solution for this problem?

Thanks and best regards!

[Nick Cox]

The format() option is provided for such problems. See the help for egen.

 
egen ident = concat(date var2), format(%tdDDNNCCYY)

[CAPM]

Thank you very much Nick. This makes my day!

[Carlo Lazzaro]

CAPM:
please, make our day, too and, as Nick pointed out, re-register using your real names (as per FAQ). Thank you.

Kind regards,
Carlo