simple m:1 merge

[Olena Onishchenko]

Dear Stata Users

I am performing simple m:1 merge (key variable is code)on two following datasets:

Master File:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float date long tradenumber str6 code double(tradetime tradeprice)
19725 560042 "1101  " 41656000  45.4
19725 517187 "1101  " 40576000 45.35
19725 358890 "1101  " 36976000  45.4
19725 743572 "1101  " 46427000 45.25
19725 344748 "1101  " 36796000  45.5
19725 626748 "1102  " 43400000 38.15
19725 197523 "1102  " 34503000  38.4
19725 346654 "1102  " 36829000  38.1
19725 613913 "1102  " 43070000  38.2
19725 530370 "1103  " 40866000  15.7
19725 547782 "1103  " 41317000 15.75
19725 371040 "1103  " 37101000  15.7
19725 331446 "070494" 36551000   .86
19725 331454 "070973" 36551000  1.38
end
format %tdDD/NN/CCYY date
format %tc_HH:MM:SS tradetime

Using file:

* Example generated by -dataex-. To install: ssc install dataex
clear
input str4 code str12 ISIN
"1101" "TW0001101004"
"1102" "TW0001102002"
"1103" "TW0001103000"
end

 merge m:1 code using "G:\CODES.dta"

Result # of obs.
-----------------------------------------
not matched 17
from master 14 (_merge==1)
from using 3 (_merge==2)

matched 0 (_merge==3)
-----------------------------------------

Is the length of the code variable a problem? Or what am I missing here?

[Barbara Voe]

Hi Olena,
it seems as if you have blank spaces in the code in the master file! Stata is sensitive to those, so make sure your codes really match precisely.
I think you can do so by using
gen new_code = substr(code, 1, 4) if substr(code, 1, 2) == "11"
replace new_code = code if new_code==.

and then match using this new code (ensuring that the master and using file have the same variable name).

[Olena Onishchenko]

Barbara

Thank you.

Your code works for the sample data I've posted. In the original dataset, however, code does not necessarily have the first two numbers =="11".

How can I change the code variable in the using file to have 2 blank spaces after 4 digits?

Regards,
Olena

[Olena Onishchenko]

Barbara

I think I have sorted my issue by removing trailing blanks:

gen code_n=strrtrim( code )

Thank you.

Regards,
Olena

[Barbara Voe]

Perfect! Glad you sorted it out!