And there are at least two ways how to avoid recoding:

1) use COUNT with the keyword sysmis

2) use COMPUTE and the function nmiss.

Both ways are shown in this example syntax:

data list list / a b c . begin data . 1 1 2 3 1 2

end data.

count nmiss1 = a b c (sysmis) . compute nmiss2 = nmiss(a,b,c). exe. form all (f1).

HTH

Jan

-----Original Message----- From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU]On Behalf Of Hector Maletta Sent: Tuesday, September 21, 2004 10:35 PM To: SPSSX-L@LISTSERV.UGA.EDU Subject: Re: Matching cases

You may recode the missing values to some non-missing value and then count it. For instance: IF (SYSMIS(var))newvar=999. Do this for all your variables, then count the value 999 among them. I use a rather odd value like 999 to make sure it is not already among the valid values. Choose a suitable value for your case.

Hector

> -----Original Message----- > From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] > On Behalf Of Victor Tarrago > Sent: Tuesday, September 21, 2004 4:37 PM > To: SPSSX-L@LISTSERV.UGA.EDU > Subject: Re: Matching cases > > > Dear Hector, > > Is it possible to count with that function how many missing > values are? > > Thank you very much for yor help. > > > Victor Tarragó Sanromà > Acziona Quanlitative Market Consultants > Tel. 00 34 93. 238.43.53 > Mòbil 00 34 630.91.68.08 > Fax. 00 34 93.238.45.99 > vtarrago@acziona.com > www.acziona.com > > > -----Mensaje original----- > De: Hector Maletta [mailto:hmaletta@fibertel.com.ar] > Enviado el: martes, 21 de septiembre de 2004 22:15 > Para: Victor Tarrago; SPSSX-L@LISTSERV.UGA.EDU > Asunto: RE: Matching cases > > Victor, > Use the COUNT function. It counts how many times a given > value appears among a set of variables. Hector > > > -----Original Message----- > > From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] > On Behalf > > Of Victor Tarrago > > Sent: Tuesday, September 21, 2004 1:13 PM > > To: SPSSX-L@LISTSERV.UGA.EDU > > Subject: Re: Matching cases > > > > > > Dear SPSS experts, > > > > We've lots of variables. These variables could be missing (no > > response) or values between 1 to 6, 8 to 10, 20 to 22, 30 > to 36 and 97 > > to 99. We need to create a new variable for each possible > value that > > counts the number of times each value apears everywhere in > the set of > > variables. > > > > Which is the fastest way to create those variables? > > > > Thank you in advance > > > > > > Victor Tarragó Sanromà > > Acziona Quanlitative Market Consultants > > Tel. 00 34 93. 238.43.53 > > Mòbil 00 34 630.91.68.08 > > Fax. 00 34 93.238.45.99 > > vtarrago@acziona.com > > www.acziona.com > > > > > > -----Mensaje original----- > > De: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] > En nombre de > > Hector Maletta Enviado el: lunes, 23 de agosto de 2004 14:12 > > Para: SPSSX-L@LISTSERV.UGA.EDU > > Asunto: Re: Matching cases > > > > Patrick: > > Matching and merging the cases, as you intend to do, can be > done by a > > combination of commands, but as far as two records are to be merged > > into one record with twice the number of variables, the > variable names > > MUST be changed because they must be unique. I propose the > following > > syntax: *This solution assumes there is maximum one > repetition of the > > same client and part codes. *If the same client buys the same part > > more than two times the solution won't work. *Identify cases > > with the same code and put them together in the file. SORT > > CASES BY CLIENT_CODE PART_CODE. COMPUTE FLAG=(CLIENT_CODE = > > LAG(CLIENT_CODE) AND PART_CODE=LAG(PART_CODE)). > > > > *Separate both sets of cases in separate files. > > TEMPORARY. > > SELECT IF (FLAG=1). > > RENAME VARIABLES (var1 var2 var3 var 4 = avar1 avar2 avar3 avar4). > > *Substitute proper variable names in the above command, for all > > variables except client code and part code). SAVE OUTFILE > > 'SECONDTIME.SAV'/drop flag. > > > > TEMPORARY. > > SELECT IF (FLAG=0). > > SAVE OUTFILE 'FIRSTTIME.SAV'/drop flag. > > > > *Merge the files. > > *This part requires there is only one repetition of the same codes. > > GET FILE 'FIRSTTIME.SAV'. MATCH FILES /FILE=*/FILE > 'SECONDTIME.SAV'/BY > > CLIENT_CODE PART_CODE. SAVE OUTFILE 'BOTHTIMES.SAV'. > > > > *In the new file, the second set of variables will be blank > > (SYSMIS) for clients who do not have repeated the same part. > > > > Hope this helps. > > > > Hector > > > > > -----Original Message----- > > > From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] > > On Behalf > > > Of Kamau, Patrick > > > Sent: Monday, August 23, 2004 8:53 AM > > > To: SPSSX-L@LISTSERV.UGA.EDU > > > Subject: Matching cases > > > > > > > > > Hi all, I need help so urgently! > > > > > > > > > > > > I have a dataset of >80,000 records which contains many > > variables two > > > of which are 'client_code' and 'part_code' (both are > string). If a > > > client comes in with a partner, s/he is given a code, say > > > 'client_code' =10111/03/0001 and on the same record, a > > partner code is > > > filled, say 'part_code' =10111/03/0002. On the same > > dataset, there has > > > to be another record with 'client_code' =10111/03/002 and > > 'part_code' > > > =10111/03/0001. > > > > > > > > > > > > I would like to match these two records and merge to one > > record with > > > double number of variables so that I can compare the > variables, and > > > for those which don't have a match, I can find out why. Example: > > > > > > client_code part_code > > > sex age > > > reason1 reason2 status > > > > > > 10111/03/0001 10111/03/0002 1 > > 23 > > > IDU no condom +VE > > > > > > 10111/03/0002 10111/03/0001 2 > > 20 > > > partner FSW +VE > > > > > > 10111/03/0320 1 > > > 19 MISS MISS -VE > > > > > > 10111/03/0305 10111/03/0340 1 > > 26 > > > MISS MISS +VE > > > > > > > > > > > > I would like this to be > > > > > > Aclient_code Bclient_code > > > Asex Bsex > > > Aage Bage Areason1 > > > Breason1................................. > > > > > > 10111/03/0001 10111/03/0002 1 > 2 > > > 23 20 IDU partner > > > > > > 10111/03/0305 10111/03/0340 1 > . > > > 26 . MISS . > > > > > > . 10111/03/0320 . > > > 1 . 19 . MISS > > > > > > > > > > > > Is there a way I can do this without renaming the variables > > of one or > > > both subsets of data (i.e. males to Avariable and females to > > > Bvariable), and if no, is there an easier way of renaming > > variables in > > > SPSS. > > > > > > > > > > > > Your help will highly be appreciated. > > > > > > > > > > > > Thank you. > > > > > > > > > > > > > > > > > > Patrick Kamau > > > > > > Senior Data Manager (VCT) > > > > > > CDC - Nairobi > > > > > > P.O. Box 606 - 00621 > > > > > > Nairobi, Kenya > > > > > > > > > > > > Tel: +254 (20) 273077(6-9) > > > > > > Fax: +254 (20) 2714745 > > > > > > Mobile: +254 (722) 778208 > > > > > > > > > > > > email:pkamau@cdcnairobi.mimcom.net > > > > > >