compute / recode variables

[Mohammed Mustafa]

Dear All,
 
i'm having a series of 21 products each product is being evaluated using 10 different parameters (210 columns) on a scale of 1-99
each data record (row) is having responses for only 1 product out of the 21 and the rest would be blank
what i need to do is to move all the products from the 210 columns to only one group of 10 columns as i'm having an indicator to know which product is being used for each respondent.
i tried compute and recode but it didn't work
can somebody please help me with that
 
Thanks in advance
 
Mohd


 

[Bruce Weaver]

Mohammed Mustafa wrote

Dear All,

i'm having a series of 21 products each product is being evaluated using 10 different parameters (210 columns) on a scale of 1-99
each data record (row) is having responses for only 1 product out of the 21 and the rest would be blank
what i need to do is to move all the products from the 210 columns to only one group of 10 columns as i'm having an indicator to know which product is being used for each respondent.
i tried compute and recode but it didn't work
can somebody please help me with that

Thanks in advance

Mohd

Does this get you close?

VARSTOCASES
  /ID=id
  /MAKE score FROM prod1param1 to prod21param10
  /INDEX=product(21) parameter(10)
  /NULL=DROP.

[Art Kendall]

What format is the original data in?
Can you cobble together an example, say, with 3 products and 3 parameters that looks like your input data as it is? Then show what you want it to look like?

Art Kendall
Social Research Consultants

On 1/4/2011 5:12 AM, Mohammed Mustafa wrote:

Dear All,
 
i'm having a series of 21 products each product is being evaluated using 10 different parameters (210 columns) on a scale of 1-99
each data record (row) is having responses for only 1 product out of the 21 and the rest would be blank
what i need to do is to move all the products from the 210 columns to only one group of 10 columns as i'm having an indicator to know which product is being used for each respondent.
i tried compute and recode but it didn't work
can somebody please help me with that
 
Thanks in advance
 
Mohd


 

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

[Nirit Avnimelech]

Assuming your data looks something like this:

 

Product                p1_s1… p1_s10                  p2_s1…p2_s10  ……… p21_s1…p21_s10

1                              scores for prod.1              empty                                   empty

2                              empty                                   scores for prod.2              empty

…

21                           empty                                   empty                                   scores for prod.21

 

How about the following:

DO REPEAT var1=p1_s1 to p1_s10 /var2=p2_s1 to p2_s10 /….   /var21=p21_s1 to p21_s10 /newvar=s1 to s10.

COMPUTE newvar=max(var1,var2,var3,….,var21).

END REPEAT.

 

The variables s1 to s10 will all be full, each of the scores corresponding to “product”.

 

Hope this helps.

 

Nirit

 

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Mohammed Mustafa
Sent: Tuesday, January 04, 2011 12:13 PM
To: [hidden email]
Subject: compute / recode variables

 

Dear All,
 
i'm having a series of 21 products each product is being evaluated using 10 different parameters (210 columns) on a scale of 1-99
each data record (row) is having responses for only 1 product out of the 21 and the rest would be blank
what i need to do is to move all the products from the 210 columns to only one group of 10 columns as i'm having an indicator to know which product is being used for each respondent.
i tried compute and recode but it didn't work
can somebody please help me with that
 
Thanks in advance
 
Mohd


 

[David Marso]

I leave it to you to study and generalize the following:
Also, in the future it would be helpful to post what you "tried but didn't work" .
HTH, David

data list free / ID x1_p1 TO x1_p4 X2_p1 To X2_p4 x3_p1 TO X3_p4.
begin data
01 11 12 13 14  .  .  .  .  .  .  .  .
02  .  .  .  . 21 22 23 24  .  .  .  .
03  .  .  .  .  .  .  .  . 31 32 33 34
end data.
VECTOR XP=X1_P1 TO X3_P4 /KEEP(4).
LOOP X=1 TO 3.
+  LOOP P=1 TO 4.
+    DO IF NOT MISSING ( XP((X-1)*4+P)) .
+      COMPUTE KEEP(P)=  XP((X-1)*4+P) .
+      COMPUTE INDIC=X.
+    END IF.
+  END LOOP.
END LOOP.
XSAVE OUTFILE "TEMP" / KEEP ID INDIC KEEP1 TO KEEP4.
EXE.
GET FILE "TEMP" .
list.


      ID    INDIC    KEEP1    KEEP2    KEEP3    KEEP4

    1.00     1.00    11.00    12.00    13.00    14.00
    2.00     2.00    21.00    22.00    23.00    24.00
    3.00     3.00    31.00    32.00    33.00    34.00


Number of cases read:  3    Number of cases listed:  3

Mohammed Mustafa wrote

Dear All,

i'm having a series of 21 products each product is being evaluated using 10 different parameters (210 columns) on a scale of 1-99
each data record (row) is having responses for only 1 product out of the 21 and the rest would be blank
what i need to do is to move all the products from the 210 columns to only one group of 10 columns as i'm having an indicator to know which product is being used for each respondent.
i tried compute and recode but it didn't work
can somebody please help me with that

Thanks in advance

Mohd

[Bruce Weaver]

Ah, okay.  If that is the desired result, my VARSTOCASES method was incomplete.  One would have to add a CASESTOVARS, like this:

data list free / ID x1_p1 TO x1_p4 X2_p1 To X2_p4 x3_p1 TO X3_p4.
begin data
01 11 12 13 14  .  .  .  .  .  .  .  .
02  .  .  .  . 21 22 23 24  .  .  .  .
03  .  .  .  .  .  .  .  . 31 32 33 34
04  .  .  .  . 21 22 23 24  .  .  .  .
05 11 12 13 14  .  .  .  .  .  .  .  .
end data.

VARSTOCASES
  /MAKE parameter FROM x1_p1 to x3_p4
  /INDEX=product(3) parameter_number(4)
  /NULL=DROP.

CASESTOVARS
  /ID=ID product
  /INDEX=parameter_number
  /GROUPBY=VARIABLE.

RESULT:

   ID product parameter.1 parameter.2 parameter.3 parameter.4

    1      1        11          12          13          14
    2      2        21          22          23          24
    3      3        31          32          33          34
    4      2        21          22          23          24
    5      1        11          12          13          14

David Marso wrote

I leave it to you to study and generalize the following:
Also, in the future it would be helpful to post what you "tried but didn't work" .
HTH, David

data list free / ID x1_p1 TO x1_p4 X2_p1 To X2_p4 x3_p1 TO X3_p4.
begin data
01 11 12 13 14  .  .  .  .  .  .  .  .
02  .  .  .  . 21 22 23 24  .  .  .  .
03  .  .  .  .  .  .  .  . 31 32 33 34
end data.
VECTOR XP=X1_P1 TO X3_P4 /KEEP(4).
LOOP X=1 TO 3.
+  LOOP P=1 TO 4.
+    DO IF NOT MISSING ( XP((X-1)*4+P)) .
+      COMPUTE KEEP(P)=  XP((X-1)*4+P) .
+      COMPUTE INDIC=X.
+    END IF.
+  END LOOP.
END LOOP.
XSAVE OUTFILE "TEMP" / KEEP ID INDIC KEEP1 TO KEEP4.
EXE.
GET FILE "TEMP" .
list.


      ID    INDIC    KEEP1    KEEP2    KEEP3    KEEP4

    1.00     1.00    11.00    12.00    13.00    14.00
    2.00     2.00    21.00    22.00    23.00    24.00
    3.00     3.00    31.00    32.00    33.00    34.00


Number of cases read:  3    Number of cases listed:  3

Mohammed Mustafa wrote

Dear All,

i'm having a series of 21 products each product is being evaluated using 10 different parameters (210 columns) on a scale of 1-99
each data record (row) is having responses for only 1 product out of the 21 and the rest would be blank
what i need to do is to move all the products from the 210 columns to only one group of 10 columns as i'm having an indicator to know which product is being used for each respondent.
i tried compute and recode but it didn't work
can somebody please help me with that

Thanks in advance

Mohd