ID variable

[Sam Ballard]

Hi All,

I have a dataset with duplicate cases and I'm trying to create a new
variable (see below). I want the new variable equal to one if VAR1 equaled
one for any of the ID's. Does anyone know a simple way to do this?

ID     var1   new
101     0     1
101     0     1
101     1     1
102     0     0
102     0     0
103     1     1
103     0     1
103     0     1

Thanks,
Sam
SDCCD

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

[Maguin, Eugene]

Sam,

There are at least two ways (and maybe more) to do this.

My usual way would be this.

Sort cases by id (a) var1 (d).
Compute new=0.
Do if ($casenum eq 1 or id ne lag(id)).
+  if (var1 eq 1) new=1.
Else.
+  if (lag(new) eq 1) new=1.
End if.


But another way that others may have suggested is this. I haven't use this
method but it should work.

Aggegate outfile=* mode=addvariables/break=id/new=fgt(var1,0).

Recode new(0=0)(else=1).


Gene Maguin




>>I have a dataset with duplicate cases and I'm trying to create a new
variable (see below). I want the new variable equal to one if VAR1 equaled
one for any of the ID's. Does anyone know a simple way to do this?

ID     var1   new
101     0     1
101     0     1
101     1     1
102     0     0
102     0     0
103     1     1
103     0     1
103     0     1

Thanks,
Sam
SDCCD

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

[Bruce Weaver]

Sam Ballard wrote

Hi All,

I have a dataset with duplicate cases and I'm trying to create a new
variable (see below). I want the new variable equal to one if VAR1 equaled
one for any of the ID's. Does anyone know a simple way to do this?

ID     var1   new
101     0     1
101     0     1
101     1     1
102     0     0
102     0     0
103     1     1
103     0     1
103     0     1

Thanks,
Sam
SDCCD

* IF 0 and 1 are the only possible values for VAR1,
* and if the cases are already sorted by ID,
* let NEW = the maximum value of VAR1 within each ID.

aggregate / break = id / new = max(var1).
list.

The output:

 ID var1 new

101    0   1
101    0   1
101    1   1
102    0   0
102    0   0
103    1   1
103    0   1
103    0   1