Removing consecutive spaces in string variable

[Divya Singh]

Dear all,

I have been struggling with removing spaces from a string variable called booth_id. Here is how it looks:

* Example generated by -dataex-. To install: ssc install dataex
clear
input str62 booth_id
"98-4     "
"98-5     "
"98-6     "
"98-7     "
"98-8     "
"98-9"     
"98-10 "   
"98-11 "   
"98-12 "   
end

I have tried removing spaces, tabs in all the possible ways I could think of but nothing works and 0 real changes are made.

replace booth_id=itrim(booth_id)
replace booth_id=ltrim(booth_id)
replace booth_id= subinstr(booth_id, char(9), "",.) 
replace booth_id= subinstr(booth_id, " ", "",.)

When I try to destring after splitting booth_id by "-", it says non-numeric characters however when I do charlist for the second part of the string (following "-"), it returns the following:
0123456789

I don't understand how to remove these spaces from booth_id if they are indeed spaces. I would be grateful if somebody could help me with this.

[Clyde Schechter]

Well, the first two won't do what you want because -itrim()- is designed to remove spaces between words: but you don't have any of those, you just have them after. Similarly, -ltrim()- won't do it because that is designed to remove leading spaces, and you only have trailing spaces. -subinstr(booth_id, char(9), "", .)- won't do it because char(9) is not the space character. The space character is char(32). But all of the following work with your example data:

replace booth_id=rtrim(booth_id)
replace booth_id = trim(booth_id)
replace booth_id= subinstr(booth_id, char(32), "",.) 
replace booth_id= subinstr(booth_id, " ", "",.)

You say you tried the last of these yourself and it didn't replace any characters, but I can't replicate that behavior on my setup. For me, it does exactly what you want.

And thank you for using -dataex-.

[Nick Cox]

Note that itrim() reduces multiple internal spaces to single spaces:

. di  itrim("This   is         a   test")
This is a test

So that's removing (all but one of) spaces between each pair of words.

[Divya Singh]

Thank you Clyde and Nick for your very valuable comments. Yes, the part of the dataset that I produced above, the code works. However there are some observations in this variable for which this doesn't work completely. Here are those:

* Example generated by -dataex-. To install: ssc install dataex
clear
input str62 booth_id
"98-114             "
"98-115             "
"98-116             "
"98-117             "
"98-118             "
"98-119             "
"98-120             "
"98-121             "
"98-122             "
"98-123             "
"98-124             "
"98-125             "
"98-126             "
"98-127             "
"98-128             "
"98-129             "
"98-130             "
"98-131             "
"98-132             "
end

[Clyde Schechter]

You need to post the exact commands you ran and show the exact results Stata gave you. When I run each of those four commands on this data in my setup, each one of them removes all of the blanks from all 19 observations:

. * Example generated by -dataex-. To install: ssc install dataex
. clear

. input str62 booth_id

                                                           booth_id
  1. "98-114             "
  2. "98-115             "
  3. "98-116             "
  4. "98-117             "
  5. "98-118             "
  6. "98-119             "
  7. "98-120             "
  8. "98-121             "
  9. "98-122             "
 10. "98-123             "
 11. "98-124             "
 12. "98-125             "
 13. "98-126             "
 14. "98-127             "
 15. "98-128             "
 16. "98-129             "
 17. "98-130             "
 18. "98-131             "
 19. "98-132             "
 20. end

. 
. preserve

. replace booth_id=rtrim(booth_id)
(19 real changes made)

. restore, preserve

. replace booth_id = trim(booth_id)
(19 real changes made)

. restore, preserve

. replace booth_id= subinstr(booth_id, char(32), "",.) 
(19 real changes made)

. restore

. replace booth_id= subinstr(booth_id, " ", "",.)
(19 real changes made)

.

So you are doing something that is somehow different, and without seeing exactly what, there is no way to advise you.

[Divya Singh]

Sure, this is what I did:

. tab booth_id

                               booth_id |      Freq.     Percent        Cum.
----------------------------------------+-----------------------------------
                    98-113              |          1        9.09        9.09
                    98-114              |          1        9.09       18.18
                    98-115              |          1        9.09       27.27
                    98-116              |          1        9.09       36.36
                    98-117              |          1        9.09       45.45
                    98-118              |          1        9.09       54.55
                    98-119              |          1        9.09       63.64
                    98-120              |          1        9.09       72.73
                    98-121              |          1        9.09       81.82
                    98-122              |          1        9.09       90.91
                    98-123              |          1        9.09      100.00
----------------------------------------+-----------------------------------
                                  Total |         11      100.00

. 
. replace booth_id=rtrim(booth_id)
(0 real changes made)

. replace booth_id = trim(booth_id)
(0 real changes made)

. replace booth_id= subinstr(booth_id, char(32), "",.) 
(11 real changes made)

. replace booth_id= subinstr(booth_id, " ", "",.)
(0 real changes made)

. tab booth_id

                               booth_id |      Freq.     Percent        Cum.
----------------------------------------+-----------------------------------
                     98-113             |          1        9.09        9.09
                     98-114             |          1        9.09       18.18
                     98-115             |          1        9.09       27.27
                     98-116             |          1        9.09       36.36
                     98-117             |          1        9.09       45.45
                     98-118             |          1        9.09       54.55
                     98-119             |          1        9.09       63.64
                     98-120             |          1        9.09       72.73
                     98-121             |          1        9.09       81.82
                     98-122             |          1        9.09       90.91
                     98-123             |          1        9.09      100.00
----------------------------------------+-----------------------------------
                                  Total |         11      100.00
. charlist booth_id
-0123456789

But I see when I copy-paste the -dateex- input, the above code works perfectly. Could it because I imported this data from excel and maybe data is not read properly? Thank you.

[Nick Cox]

The help for charlist (SSC, as you are asked to explain)discusses what to do with unprintable and/or invisible codes, such as char(160).

[Divya Singh]

Thank you so much for your help, Nick and Clyde. Apparently these were not spaces but some invisible characters. In the end, I solved this problem by opening the data in another editor and removing them.

[Nick Cox]

Thanks for reporting and good that you solved your problem. You could have removed them in Stata too: the help for charlist gives examples.