destring non numeric values

[hatem ali]

Hello
I am Hatem Ali, STATA user.
I have a 2 large dataset that I am trying to merge
The common variable in these 2 datasets is called: trr-id_code.
However, This variable is a string non-numeric variable

When I try to merge one by one using key variable. STATA doesnt do and say variables are unequal.

when I tried to destring it, I got the following message:
destring trr_id_code, generate(hhh)
trr_id_code contains nonnumeric characters; no generate

1st question: Is there a way I can destring this variable so I can use it for merging?
2nd question: can I use : merge one to one by observation instead?

Looking forward to hear back from you

[Carlo Lazzaro]

Hatem:
if you have an unique observation for each -trr-id_code-, you may want to try:

. g trr_id_code="A" in 1
(1 missing value generated)

. replace trr_id_code="B" in 2
(1 real change made)

. egen id_num=group( trr_id_code)

. list

     +-------------------+
     | trr_id~e   id_num |
     |-------------------|
  1. |        A        1 |
  2. |        B        2 |
     +-------------------+

and then try -merge-.

As an aside, things would be easier to inspect if you shared an excerpt/example of your data via -dataex-.

[Rich Goldstein]

first, you can merge on string variables; your report on the message from Stata is unclear, please post exactly what Stata said (and put it in CODE blocks - see the FAQ)

second, -destring- is for variables that are really numeric but happen to be string due to the way they were imported or input; -encode- for making numeric variables out of string variables

third, if you really want to merge by observation, go into each data set and

gen newid=_n
sort newid

and then -merge- on "newid"

see

help merge
help encode

[hatem ali]

Thanks Rich and Carlo for getting back to me.

This is an example of the data. The variable :trr_id_code is a string variable.
When I try to merge on it using one by one key code, the following message appear: variable trr_id_code does not uniquely identify observations in the master data

This message still appear even after removing the duplicates

trr_id_code	dx_date_lymph	pathology	steroids_maint
A100140	19/07/1999	1	1
A100440	29/10/2014	0	1
A100980	12/07/2013	1	1
A102013	02/11/2008	0	1
A102130	13/01/2001	1
A103307	01/01/2000	1	0

This is example of second dataset

trr_id_code	alg_ind	alg_days
A100	1	4
A100001	1	5
A100002	1	6
A100005	0	0
A100006	0	0
A100007	0	0

[hatem ali]

This is the dexription and summary of the varname

storage display value
variable name type format label variable label

trr_id_code str16 %16s ENCRYPTED TRR_ID



. summarize trr_id_code

Variable Obs Mean Std. Dev. Min Max

trr_id_code 0

[hatem ali]

Thanks all,

Got it , done