Droping "NA"s

[Pedro Coelho]

Hi Stata users.

I have 26 variables that contain some observations as NA. Is there a way (

olther than "

drop if var==NA" for each variable)

to drop all observations that contain "NA"s for any variable?

* Example generated by -dataex-. To install: ssc install dataex
clear
input double id str17 mtb0 str9(ebv0 debt0 ebit0 ta1 ni1) str8 cfo1
302319 "NA"       "-1364"    "376"     "-1055"   "117"      "-476"    "-278"  
303063 ".83"      "-544"     "952"     "-2402"   "1486"     "-667"    "-1365" 
282874 "238.15"   "200229"   "32358"   "35823"   "260814"   "30219"   "1883"  
214440 "261.88"   "164534"   "0"       "6766"    "166933"   "6752"    "-3199" 
378443 "1394.28"  "2887794"  "0"       "-760338" "11451238" "-391191" "350907"
301631 "81305.81" "16763000" "7864000" "7905000" "35994000" "5142000" "6999000"
283489 "6.9"      "NA"       "NA"      "NA"      "9699"     "-21888"  "-9224" 
end

Thanks in advance.

[Iuri Gavronski]

I had a similar problem (importing a CSV file that otherwise would work on R...) and I solved with a loop:

foreach var of varlist * {
    replace `var'="" if `var'=="NA"
    }

My code replaced NAs with missing. I believe you can adjust it to either drop the observation or have the procedure (say, -reg-) do it for you.

Best,

Iuri.

[Pedro Coelho]

Thanks for your suggestion Iuri. I also used this loop but somehow didn't manage to do what I need.

[Carlo Lazzaro]

Pedro:
what follows will do the trick:

. foreach var of varlist mtb0-cfo1 {
  2. drop if `var'=="NA"
  3.  }
(1 observation deleted)
(1 observation deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)
(0 observations deleted)

. list

     +---------------------------------------------------------------------------------+
     |     id       mtb0       ebv0     debt0     ebit0        ta1       ni1      cfo1 |
     |---------------------------------------------------------------------------------|
  1. | 303063        .83       -544       952     -2402       1486      -667     -1365 |
  2. | 282874     238.15     200229     32358     35823     260814     30219      1883 |
  3. | 214440     261.88     164534         0      6766     166933      6752     -3199 |
  4. | 378443    1394.28    2887794         0   -760338   11451238   -391191    350907 |
  5. | 301631   81305.81   16763000   7864000   7905000   35994000   5142000   6999000 |
     +---------------------------------------------------------------------------------+
.

[Pedro Coelho]

Thank you Carlo, works just perfect!
That's probably the adjust Iuri suggested but I didn't manage to find.

Thank you both.

[Carlo Lazzaro]

Pedro:
happy with reading that you solved your problem.
For the future, when something (command, code, else) does not produce what you expected it to, please act on https://www.statalist.org/forums/help#technical , especially #12.1.
If you follow that advice, interested listers can spot faster what's the matter with your command, code or else and reply positively. Thanks.

[Pedro Coelho]

Thanks for your advice, Carlo. I'll certainly follow it.

Best regards.