More efficient code?

[parisec]

Hi all,

I'm computing a new variable x2 with 9 categories from x1 which is 122 categories, my syntax looks like this:

compute x2=$sysmis.
    (if x1=2 or x1=6 or x1=65 or x1=68 or x1=79) x2=1.
    (if x1 = 3 or x1 = 7 or x1=10 or x1 = 12) x2 = 2.
    ........etc

This works just fine. But does more efficient syntax exist for something like this? Some of these new x2 categories include 40 or so categories of x1 making for a high probability of screwing up.

It's a new year and I should learn some new syntax.

Thanks much.
Carol

[Andy W]

When I get that many categories (and the order is irregular) I don't bother with IF or whatever, but actually make a separate look up table to define the second category. E.g. I would have a table that maps x1 to x2;

X1 X2
 2 1
 6 1
65 1
68 1
79 1
 3 2
 7 2
10 2
12 2
etc....

Then having this as a separate SPSS data file,  the code is simply:

SORT CASES BY x1.
MATCH FILES FILE = *
  /TABLE = 'LookUpTable'
  /BY x1.

[Jon K Peck]

Makes sense for big category lists, but the any function would also simplify the posted code a lot.


Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
phone: [phone]




From:        Andy W <[hidden email]>
To:        [hidden email]
Date:        01/02/2015 10:25 AM
Subject:        Re: [SPSSX-L] More efficient code?
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

When I get that many categories (and the order is irregular) I don't bother
with IF or whatever, but actually make a separate look up table to define
the second category. E.g. I would have a table that maps x1 to x2;

X1 X2
2 1
6 1
65 1
68 1
79 1
3 2
7 2
10 2
12 2
etc....

Then having this as a separate SPSS data file,  the code is simply:

SORT CASES BY x1.
MATCH FILES FILE = *
 /TABLE = 'LookUpTable'
 /BY x1.






-----
Andy W
[hidden email]
http://andrewpwheeler.wordpress.com/
--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/More-efficient-code-tp5728321p5728322.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
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


===================== 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

[David Marso]

You could go with a LOOKUP table as Andy suggested.
Or see ANY function and RECODE.

COMPUTE x2=SUM(ANY(x1, 2, 6, 65, 68, 79)*1,
                            ANY(x1, 3, 7, 10, 12 )*2) ......etc.

or
RECODE x1 (2 ,6 ,65 ,68 ,79 =1)(3,7,10,12=2) INTO x2 .

Of these 2 the RECODE is surely more efficient.

----

parisec wrote

Hi all,

I'm computing a new variable x2 with 9 categories from x1 which is 122 categories, my syntax looks like this:

compute x2=$sysmis.
    (if x1=2 or x1=6 or x1=65 or x1=68 or x1=79) x2=1.
    (if x1 = 3 or x1 = 7 or x1=10 or x1 = 12) x2 = 2.
    ........etc

This works just fine. But does more efficient syntax exist for something like this? Some of these new x2 categories include 40 or so categories of x1 making for a high probability of screwing up.

It's a new year and I should learn some new syntax.

Thanks much.
Carol

[parisec]

thanks all!  


Now i have three methods that are way better than what i have been doing.

Great way to start the year.

Carol