topic Re: Here are the values in thsoe tow vars in SAS Programming

What is wrong with this code?

Dhana18 — Thu, 23 Jan 2020 17:45:12 GMT

Good morning,
the new vari i am trying to create MENSEX did not create. What is wrong with this code below?

DATA NovDec19.MERGED_NUMPARTS_1;
length mensex $3;
SET NovDec19.MERGED_NovDec19_NUMPARTS;
if gender_sp ="1" AND NUM_PRT_3M="1" THEN MENSEX= "1";
if gender_sp ="1" AND NUM_PRT_3M="2" THEN MENSEX="2";
if gender_sp ="1" AND NUM_PRT_3M="3" THEN MENSEX="3";
if gender_sp ="1" AND NUM_PRT_3M="5" THEN MENSEX="5";
if gender_sp in("3" , "9") AND NUM_PRT_3M in("1" , "2" , "3" , "5") THEN MENSEX="999";
run;

Re: What is wrong with this code?

ballardw — Thu, 23 Jan 2020 15:21:43 GMT

Did you look at the log after running the code? If there were any notes about variables not found, not referenced or data conversions that could be one cause. Post a copy of the log from running the if there are any messages of such type.

Otherwise we would need example data. Instructions here: https://communities.sas.com/t5/SAS-Communities-Library/How-to-create-a-data-step-version-of-your-data-AKA-generate/ta-p/258712 will show how to turn an existing SAS data set into data step code that can be pasted into a forum code box using the {i} icon or attached as text to show exactly what you have and that we can test code against.

The example data only needs to include the variables you show being used in your code.

Re: What is wrong with this code?

PaigeMiller — Thu, 23 Jan 2020 15:22:19 GMT

Since we don't have the data set NovDec19.MERGED_NovDec19_NUMPARTS, you need to look at this data set and see what values are present for variables gender_sp and NUM_PRT_3M.

Also, please look at the log for ERROR or WARNING messages.

Re: What is wrong with this code?

Dhana18 — Thu, 23 Jan 2020 15:36:43 GMT

Here is the log:
DATA NovDec19.MERGED_NUMPARTS_1;
3 length mensex $3;
4 SET NovDec19.MERGED_NovDec19_NUMPARTS;
5 if gender_sp ="1" AND NUM_PRT_3M="1" THEN MENSEX= "1";
6 if gender_sp ="1" AND NUM_PRT_3M="2" THEN MENSEX="2";
7 if gender_sp ="1" AND NUM_PRT_3M="3" THEN MENSEX="3";
8 if gender_sp ="1" AND NUM_PRT_3M="5" THEN MENSEX="5";
9 if gender_sp in("3" , "9") AND NUM_PRT_3M in("1" , "2" , "3" , "5") THEN MENSEX="999";
10 run;

NOTE: There were 995 observations read from the data set NOVDEC19.MERGED_NOVDEC19_NUMPARTS.
NOTE: The data set NOVDEC19.MERGED_NUMPARTS_1 has 995 observations and 203 variables.
NOTE: DATA statement used (Total process time):
real time 0.59 seconds
cpu time 0.12 seconds

Here are the values in thsoe tow vars

Dhana18 — Thu, 23 Jan 2020 15:39:43 GMT

Table of NUM_PRT_3M by gender_sp

NUM_PRT_3M

gender_sp(gender_sp)

1

2

3

9

Total

0

128

193

17

44

382

1

128

144

4

3

279

2

70

107

13

3

193

3

8

57

11

0

76

5

16

40

6

3

65

Total

350

541

51

53

995

Re: Here are the values in thsoe tow vars

PaigeMiller — Thu, 23 Jan 2020 15:47:39 GMT

Yes, I see the numbers, but I wanted you to look at the the actual SAS dataset used (not me look at something you typed) and see if the values in each row of your dataset should produce the values of MENSEX based upon the code you showed us. I cannot tell, because I don't have the data set, you do.

Re: Here are the values in thsoe tow vars

ballardw — Thu, 23 Jan 2020 16:07:44 GMT

That 'table' isn't legible.

Try running this code on your data:

proc freq data= NovDec19.MERGED_NUMPARTS_1;
   tables  gender_sp* NUM_PRT_3M*MENSEX / missing list nopercent nocum;
run;

This will show all of the combinations of the three variables present in your data, include MISSING values, in an easier to read row layout. So you may find combination that you did not consider for your recoding.

Re: Here are the values in thsoe tow vars

Dhana18 — Thu, 23 Jan 2020 16:18:43 GMT

gender_sp NUM_PRT_3M mensex Frequency 1 0 999 128 1 1 999 128 1 2 999 70 1 3 999 8 1 5 999 16 2 0 999 193 2 1 999 144 2 2 999 107 2 3 999 57 2 5 999 40 3 0 999 17 3 1 999 4 3 2 999 13 3 3 999 11 3 5 999 6 9 0 999 44 9 1 999 3 9 2 999 3 9 5 999 3 I want this to be like this gender_sp NUM_PRT_3M mensex Frequency 1 0 999 128 1 1 1 128 1 2 1 70 1 3 1 8 1 5 1 16 2 0 999 193 2 1 1 144 2 2 2 107 2 3 3 57 2 5 5 40 3 0 999 17 3 1 1 4 3 2 2 13 3 3 3 11 3 5 5 6 9 0 999 44 9 1 1 3 9 2 2 3 9 5 5 3

Re: Here are the values in thsoe tow vars

Dhana18 — Thu, 23 Jan 2020 16:20:43 GMT

gender_sp	NUM_PRT_3M	mensex	Frequency
1	0	999	128
1	1	999	128
1	2	999	70
1	3	999	8
1	5	999	16
2	0	999	193
2	1	999	144
2	2	999	107
2	3	999	57
2	5	999	40
3	0	999	17
3	1	999	4
3	2	999	13
3	3	999	11
3	5	999	6
9	0	999	44
9	1	999	3
9	2	999	3
9	5	999	3

Re: Here are the values in thsoe tow vars

PaigeMiller — Thu, 23 Jan 2020 16:29:40 GMT

Are the variables GENDER_SP and NUM_PRT_3M formatted somehow?

Please show us a PROC CONTENTS of this data set (well, we really need to see just those two variables from the PROC CONTENTS output)

Re: Here are the values in thsoe tow vars

Dhana18 — Thu, 23 Jan 2020 16:36:42 GMT

here the proc contents result on these two vars.

Alphabetic List of Variables and Attributes
Variable	Type	Len	Format	Informat	Label
NUM_PRT_3M	Char	1
gender_sp	Char	2			gender_sp

I renamed gendersp var to gender_sp. is that what?

Re: Here are the values in thsoe tow vars

Dhana18 — Thu, 23 Jan 2020 17:24:43 GMT

Alphabetic List of Variables and Attributes

Variable

Type

Len

Format

Informat

Label

NUM_PRT_3M

Char

1

gender_sp

Char

2

gender_sp

Re: Here are the values in thsoe tow vars

PaigeMiller — Thu, 23 Jan 2020 17:43:22 GMT

As requested earlier by @ballardw , you need to provide us with a portion of this data set, using the instructions in the link he provided. We do not want screen captures, we do not want Excel files or attachments, we need to see a portion of this data in that format and not any other format.

Re: Here are the values in thsoe tow vars

Tom — Thu, 23 Jan 2020 17:49:50 GMT

Since GENDER_SP is length $2 it might be that the single digit codes are in the second character instead of the first. Or that the second character has some other non-graphic character other than a space.

Since it looks like the test you are making are non-overlapping it might make sense to do them as a series of ELSE clauses. Then you can include a catch all for any cases that do not fit.

DATA NovDec19.MERGED_NUMPARTS_1;
  SET NovDec19.MERGED_NovDec19_NUMPARTS;
  length mensex $3;
  if gender_sp ="1" AND NUM_PRT_3M="1" THEN MENSEX= "1";
  else if left(gender_sp) ="1" AND NUM_PRT_3M="2" THEN MENSEX="2";
  else if left(gender_sp)="1" AND NUM_PRT_3M="3" THEN MENSEX="3";
  else if left(gender_sp)="1" AND NUM_PRT_3M="5" THEN MENSEX="5";
  else if left(gender_sp) in ("3" , "9") AND NUM_PRT_3M in ("1" , "2" , "3" , "5") THEN MENSEX="999";
  else MENSEX='XXX';
run;

And most import make sure you are checking the NEW dataset and not the old one when check whether the re-code worked.

Re: What is wrong with this code?

Dhana18 — Thu, 23 Jan 2020 20:23:06 GMT

Hello, than you for your email. the instructions is too comlicated forme. Would be nic eot have a simple instructions. sorry it was not so helpful for me this time.

Re: Here are the values in thsoe tow vars

ballardw — Thu, 23 Jan 2020 21:06:23 GMT

@Dhana18 wrote:

gender_sp NUM_PRT_3M mensex Frequency

1 0 999 128

1 1 999 128

1 2 999 70

1 3 999 8

1 5 999 16

2 0 999 193

2 1 999 144

2 2 999 107

2 3 999 57

2 5 999 40

3 0 999 17

3 1 999 4

3 2 999 13

3 3 999 11

3 5 999 6

9 0 999 44

9 1 999 3

9 2 999 3

9 5 999 3

One thing that we learn eventually about SAS is that it tries to help us make "nice" output. Sometimes that interferes with diagnosing problems when using output from procedures to examine values. One of things that happens is that most procedures will strip leading spaces from values. So it is possible that the 1, 2, 3 and 9 above are actually " 1", " 2", " 3" and " 9" for the gender_sp variable so comparisons of equality such as ="1" fail.

Here is a brief example you can run. Note that the row heading for Y in the second table now is indented because the display Style instruction says to use the leading space:

data example;
   x='2';
   y=' 2';
run;

proc tabulate data=example;
title "Default display removes leading blanks";
   class x y;
   table x y , n;

run;title;
proc tabulate data=example;
title "ASIS option shows leading blanks";
   class x y;
   classlev y / style=[Asis=on];
   table x y , n;

run;title;

Which is one of the reasons that we ask about actual data an not proc output.

responses

Dhana18 — Tue, 28 Jan 2020 19:25:20 GMT

Hi All,

I am not getting answers to my problem questions as I did in the past. Did something change in the system? Am I not asking question the right way? Did I do something wrong? Do people not want to help me?

Please help me I need your help to do my job well.

thank you.

Re: responses

Tom — Tue, 28 Jan 2020 19:32:57 GMT

@Dhana18 wrote:

Hi All,

I am not getting answers to my problem questions as I did in the past. Did something change in the system? Am I not asking question the right way? Did I do something wrong? Do people not want to help me?

Please help me I need your help to do my job well.

thank you.

Are you referring to this thread? Did you try the things suggested? Did they make a difference?

Note that if you do not provide a minimal complete set of data and code that re-creates the problem then all you get is suggestions, since the people replying will have nothing to test their suggestions against.

Re: responses

PaigeMiller — Tue, 28 Jan 2020 19:41:27 GMT

As stated by @ballardw: "One thing that we learn eventually about SAS is that it tries to help us make "nice" output. Sometimes that interferes with diagnosing problems when using output from procedures to examine values." You need to provide us with the data set, in the format that we have requested, and then we ought to be able to provide help. That seems to be the only path forward — showing us the data in other forms, as you have done, which can disguise the true cause of the matter, doesn't lead to any obvious answers.

Re: Here are the values in thsoe tow vars

ballardw — Tue, 28 Jan 2020 22:12:20 GMT

@Dhana18 wrote:

gender_sp NUM_PRT_3M mensex Frequency

1 0 999 128

1 1 999 128

1 2 999 70

1 3 999 8

1 5 999 16

2 0 999 193

2 1 999 144

2 2 999 107

2 3 999 57

2 5 999 40

3 0 999 17

3 1 999 4

3 2 999 13

3 3 999 11

3 5 999 6

9 0 999 44

9 1 999 3

9 2 999 3

9 5 999 3

You are showing values of gender_sp and Num_prt_3m not in the code with results.

So I have to assume that you have some other code affecting your MENSEX variable.

Here is an example data set that just uses character values of those shown in your freq output above:

data example;
   length gender_sp $ 1 Num_prt_3m $ 2 mensex $ 3;
   do gender_sp='1','2','3','9';
      do Num_prt_3m = '0','1','2','3','5';
         if gender_sp ="1" AND NUM_PRT_3M="1" THEN MENSEX="1";
         if gender_sp ="1" AND NUM_PRT_3M="2" THEN MENSEX="2";
         if gender_sp ="1" AND NUM_PRT_3M="3" THEN MENSEX="3";
         if gender_sp ="1" AND NUM_PRT_3M="5" THEN MENSEX="5";
         if gender_sp in("3" , "9") AND NUM_PRT_3M in("1" "2" "3" "5") THEN MENSEX="999";
         output;
         /* reset result before next loop iteration*/
         call missing(mensex);
      end;
   end;
run;

Note that we get every possible combination of the one character length values of the two input variables.

And the result of:

proc freq data=example;
  tables gender_sp*Num_prt_3m*mensex/missing list nopercent nocum;
run;

The FREQ Procedure

gender_sp	Num_prt_3m	mensex	Frequency
1	0		1
1	1	1	1
1	2	2	1
1	3	3	1
1	5	5	1
2	0		1
2	1		1
2	2		1
2	3		1
2	5		1
3	0		1
3	1	999	1
3	2	999	1
3	3	999	1
3	5	999	1
9	0		1
9	1	999	1
9	2	999	1
9	3	999	1
9	5	999	1

Which does not have some of the output values you show for the same input. So either 1) your input is different than implied by the Proc Freq output and/or 2) your code as executed includes something you have not shown.

Such as a value set in a previous data step, read from the source file, or manipulated after assignment in this data step.

I suggest that you carefully search your code for every instance of MENSEX and see if you have another assignment or possibly and array using that variable.

I had removed the commas in the last comparison just incase your pasted code had contained a non-visible character that was causing problems.

gender_sp	Num_prt_3m	mensex	Frequency
1	0		1
1	1	1	1
1	2	2	1
1	3	3	1
1	5	5	1
2	0		1
2	1		1
2	2		1
2	3		1
2	5		1
3	0		1
3	1	999	1
3	2	999	1
3	3	999	1
3	5	999	1
9	0		1
9	1	999	1
9	2	999	1
9	3	999	1
9	5	999	1

gender_sp	Num_prt_3m	mensex	Frequency
1	0		1
1	1	1	1
1	2	2	1
1	3	3	1
1	5	5	1
2	0		1
2	1		1
2	2		1
2	3		1
2	5		1
3	0		1
3	1	999	1
3	2	999	1
3	3	999	1
3	5	999	1
9	0		1
9	1	999	1
9	2	999	1
9	3	999	1
9	5	999	1

gender_sp	Num_prt_3m	mensex	Frequency
1	0		1
1	1	1	1
1	2	2	1
1	3	3	1
1	5	5	1
2	0		1
2	1		1
2	2		1
2	3		1
2	5		1
3	0		1
3	1	999	1
3	2	999	1
3	3	999	1
3	5	999	1
9	0		1
9	1	999	1
9	2	999	1
9	3	999	1
9	5	999	1