casenum

[DEBOER]

Hi,
I'm trying to insert an index in my new data sample. I would like to give a unik Id to each case in a variable.

ex.

v1         id
lon         1
lon         1
lon         1
ska        2
ska        2
sto        3
sto        3
xxx       4
xxx       4

how can i create id?

thanks

[Bruce Weaver]

Look up the RANK command, and rank on v1.  Notice that there are several ways to deal with ties.

DEBOER wrote

Hi,
I'm trying to insert an index in my new data sample. I would like to give a unik Id to each case in a variable.

ex.

v1         id
lon         1
lon         1
lon         1
ska        2
ska        2
sto        3
sto        3
xxx       4
xxx       4

how can i create id?

thanks

[Mark Webb-5]

Read up on
transform --> auto-recode in the manual.

Mark Webb

Line [phone]
Cell [phone]
Fax to email [phone]
Skype  webbmark
Email  [hidden email]

On 2011/07/11 09:22 AM, DEBOER wrote:

> Hi,
> I'm trying to insert an index in my new data sample. I would like to give a
> unik Id to each case in a variable.
>
> ex.
>
> v1         id
> lon         1
> lon         1
> lon         1
> ska        2
> ska        2
> sto        3
> sto        3
> xxx       4
> xxx       4
>
> how can i create id?
>
> thanks
>
> --
> View this message in context: http://spssx-discussion.1045642.n5.nabble.com/casenum-tp4574132p4574132.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]

V1 is a string ;-)
--
Simplest way is a 2 liner..
IF (V1 NE LAG(V1)) #ID=#ID+1.
COMPUTE ID=#ID.
* If you wish to achieve deeper comprehension of how/why this works consult the manual
  and refer to SCRATCH variables.
--

Bruce Weaver wrote

Look up the RANK command, and rank on v1.  Notice that there are several ways to deal with ties.

DEBOER wrote

Hi,
I'm trying to insert an index in my new data sample. I would like to give a unik Id to each case in a variable.

ex.

v1         id
lon         1
lon         1
lon         1
ska        2
ska        2
sto        3
sto        3
xxx       4
xxx       4

how can i create id?

thanks

[Bruce Weaver]

Oops...I forgot that RANK only works on numeric variables.  If one just wants ID numbers from 1 to N though, I think Mark's AUTORECODE method is even simpler than the LAG method.  It's a one-liner, and remains so even if the cases are not sorted on V1.

new file.
dataset close all.
data list list / v1 (a3) .
begin data
lon        
lon        
ska        
ska        
lon        
sto        
xxx      
sto        
xxx      
end data.

AUTORECODE VARIABLES=v1 / INTO ID1 .

* Sort needed for LAG method.
sort cases by v1.
IF (V1 NE LAG(V1)) #ID=#ID+1.
COMPUTE ID2=#ID.

list.

David Marso wrote

V1 is a string ;-)
--
Simplest way is a 2 liner..
IF (V1 NE LAG(V1)) #ID=#ID+1.
COMPUTE ID=#ID.
* If you wish to achieve deeper comprehension of how/why this works consult the manual
  and refer to SCRATCH variables.
--

Bruce Weaver wrote

Look up the RANK command, and rank on v1.  Notice that there are several ways to deal with ties.

DEBOER wrote

Hi,
I'm trying to insert an index in my new data sample. I would like to give a unik Id to each case in a variable.

ex.

v1         id
lon         1
lon         1
lon         1
ska        2
ska        2
sto        3
sto        3
xxx       4
xxx       4

how can i create id?

thanks

[David Marso]

Point taken and well understood (my assumption was to build the values based on the sequence of cases in the file).  OTOH, if one wishes to build IDs to identify blocks of contiguous cases based on another variable irrespective of the values of other cases in the file, AUTORECODE will impose it's own ID values based on the sort order of these values.
var AR LG
aaa 1 1
aaa 1 1
ccc 4 2
ccc 4  2
bbb 3 3
bbb 3 3
xxx 6 4
uuu 5 5
uuu 5 5
aa2 2 6

Bruce Weaver wrote

Oops...I forgot that RANK only works on numeric variables.  If one just wants ID numbers from 1 to N though, I think Mark's AUTORECODE method is even simpler than the LAG method.  It's a one-liner, and remains so even if the cases are not sorted on V1.

new file.
dataset close all.
data list list / v1 (a3) .
begin data
lon        
lon        
ska        
ska        
lon        
sto        
xxx      
sto        
xxx      
end data.

AUTORECODE VARIABLES=v1 / INTO ID1 .

* Sort needed for LAG method.
sort cases by v1.
IF (V1 NE LAG(V1)) #ID=#ID+1.
COMPUTE ID2=#ID.

list.

David Marso wrote

V1 is a string ;-)
--
Simplest way is a 2 liner..
IF (V1 NE LAG(V1)) #ID=#ID+1.
COMPUTE ID=#ID.
* If you wish to achieve deeper comprehension of how/why this works consult the manual
  and refer to SCRATCH variables.
--

Bruce Weaver wrote

Look up the RANK command, and rank on v1.  Notice that there are several ways to deal with ties.

DEBOER wrote

Hi,
I'm trying to insert an index in my new data sample. I would like to give a unik Id to each case in a variable.

ex.

v1         id
lon         1
lon         1
lon         1
ska        2
ska        2
sto        3
sto        3
xxx       4
xxx       4

how can i create id?

thanks

[DEBOER]

Tanks guys, the autorecode command works perfekt