create a new data set by 'SET and 'IF'

Alexxxxxxx — Thu, 28 Feb 2019 14:27:52 GMT

Dear all

I am using the following codes to create a new sample

data Patstat_gb_hrm_set;
set Dataset.Patstat_hrm_gb_ori;
if SECTOR eq "COMPANY" or "COMPANY GOV NON-PROFIT" or "COMPANY GOV NON-PROFIT UNIVERSITY" or "COMPANY HOSPITAL" or "COMPANY UNIVERSITY";
run;

however, the result is

60   data Step1.Patstat_gb_hrm_set;
61   set Step1.Patstat_gb_hrm_set;
62   if SECTOR = 'COMPANY' or 'COMPANY GOV NON-PROFIT' or 'COMPANY GOV NON-PROFIT UNIVERSITY' or
62 ! 'COMPANY HOSPITAL' or 'COMPANY UNIVERSITY';
63   run;

NOTE: Character values have been converted to numeric values at the places given by: (Line):(Column).
      62:26    62:54    62:93    62:115
NOTE: Invalid numeric data, 'COMPANY GOV NON-PROFIT' , at line 62 column 26.
NOTE: Invalid numeric data, 'COMPANY GOV NON-PROFIT UNIVERSITY' , at line 62 column 54.
NOTE: Invalid numeric data, 'COMPANY HOSPITAL' , at line 62 column 93.
NOTE: Invalid numeric data, 'COMPANY UNIVERSITY' , at line 62 column 115.
HRM_L2=20C citation=0 earliest_filing_month=JUN2006 application=1 PERSON_CTRY_CODE=GB PERSON_NAME=20C
HRM_LEVEL=1 SECTOR=UNKNOWN PERSON_ADDRESS=Bath & Northeast Somerset _ERROR_=1 _N_=77
NOTE: Invalid numeric data, 'COMPANY GOV NON-PROFIT' , at line 62 column 26.
NOTE: Invalid numeric data, 'COMPANY GOV NON-PROFIT UNIVERSITY' , at line 62 column 54.
NOTE: Invalid numeric data, 'COMPANY HOSPITAL' , at line 62 column 93.
NOTE: Invalid numeric data, 'COMPANY UNIVERSITY' , at line 62 column 115.
....
WARNING: Limit set by ERRORS= option reached.  Further errors of this type will not be printed.
HRM_L2=A.J.N. * DE SOUSA citation=0 earliest_filing_month=FEB1992 application=1 PERSON_CTRY_CODE=GB
PERSON_NAME=A.J.N. * DE SOUSA HRM_LEVEL=1 SECTOR=UNKNOWN
PERSON_ADDRESS=31 HILLFIELD ROAD,COMBERTON,CAMBRIDGE,CB3 7DB _ERROR_=1 _N_=696
NOTE: There were 348020 observations read from the data set DATASET.PATSTAT_HRM_GB_ORI.
NOTE: The data set WORK.PATSTAT_GB_HRM_SET has 113423 observations and 9 variables.
NOTE: DATA statement used (Total process time):
      real time           13.89 seconds
      cpu time            1.39 seconds

Could you please give me some suggestion about this ? especially why Character values been converted to numeric values'?
thanks for your attention to this matter

Re: create a new data set by 'SET and 'IF'

novinosrin — Thu, 28 Feb 2019 14:30:32 GMT

Try this

data Patstat_gb_hrm_set;
set Dataset.Patstat_hrm_gb_ori;
/*if SECTOR eq "COMPANY" or "COMPANY GOV NON-PROFIT" or "COMPANY GOV NON-PROFIT UNIVERSITY" or "COMPANY HOSPITAL" or "COMPANY UNIVERSITY";*/
if SECTOR in ("COMPANY" "COMPANY GOV NON-PROFIT"  "COMPANY GOV NON-PROFIT UNIVERSITY"  "COMPANY HOSPITAL"  "COMPANY UNIVERSITY");
run;

Re: create a new data set by 'SET and 'IF'

RW9 — Thu, 28 Feb 2019 14:46:30 GMT

The if statement requires:

if <condition> then ...

So logically you are saying:

if variable sector equals the string "COMPANY" is true

or the string "COMPANY GOV NON-PROFIT" is true (which is it because its not 0)

etc.

To use if like that you need to put variable and comparitor each time:

if sector="COMPANY" or sector="COMPANY GOV NON-PROFIT" or...

topic Re: create a new data set by 'SET and 'IF' in SAS Programming

create a new data set by 'SET and 'IF'

Re: create a new data set by 'SET and 'IF'

Re: create a new data set by 'SET and 'IF'