creating custom attribute for original order.

[Art Kendall]

Sometimes I need to group string or nominal variables, so I use SORT
VARIABLES.  
After doing something like AUTORECODE on the string variables, I need to put
the variables back in the original order.
I tried creating a custom attribute. With something like Pos01 to Pos93 or
#01 to #93 or 1 to 93 as the values. That way I could sort on the custom
variable.

VARIABLE ATTRIBUTE
    VARIABLES =  ALL
    ATTRIBUTE = Input.Order('Pos') arrayname[93].

Variable Attribute No value was specified for attribute "arrayname".

With only 93 variables it is probably faster to type the contents in,
However, I thought others might need something like this, so I am posting.



-----
Art Kendall
Social Research Consultants
--
Sent from: http://spssx-discussion.1045642.n5.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

[Jon Peck]

If you use Data > Sort Variables and check the box Save the current order in a new attribute, you can recover the original order using that attribute by using it in a subsequent sort, e.g.,

SORT VARIABLES BY ATTRIBUTE oldorder (A).

On Fri, Aug 7, 2020 at 9:51 AM Art Kendall <[hidden email]> wrote:

Sometimes I need to group string or nominal variables, so I use SORT
VARIABLES. 
After doing something like AUTORECODE on the string variables, I need to put
the variables back in the original order.
I tried creating a custom attribute. With something like Pos01 to Pos93 or
#01 to #93 or 1 to 93 as the values. That way I could sort on the custom
variable.

VARIABLE ATTRIBUTE
    VARIABLES =  ALL
    ATTRIBUTE = Input.Order('Pos') arrayname[93].

Variable Attribute No value was specified for attribute "arrayname".

With only 93 variables it is probably faster to type the contents in,
However, I thought others might need something like this, so I am posting.



-----
Art Kendall
Social Research Consultants
--
Sent from: http://spssx-discussion.1045642.n5.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

[Art Kendall]

thank you.

it is interesting that FORMAT is the TYPE as a string and WIDTH also as a
string.

Followers of this list are probably aware that I emphasize complete data
definition before any analysis.

I hit my self on the head when I just now realized that sequentially using
SORT VARIABLES for LABEL,   VALUES, or  MISSING
would be a great help in this quality assurance part of the process.





-----
Art Kendall
Social Research Consultants
--
Sent from: http://spssx-discussion.1045642.n5.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

[Jon Peck]

If you want to record the original order for later use with SORT VARIABLES, this will do it.

begin program python3.
import spss
for i in range(spss.GetVariableCount()):
   spss.Submit("""VARIABLE ATTRIBUTE VARIABLES = %s ATTRIBUTE=OrigOrder('%04d').""" %\
      (spss.GetVariableName(i), i))
end program.

On Fri, Aug 7, 2020 at 9:51 AM Art Kendall <[hidden email]> wrote:

Sometimes I need to group string or nominal variables, so I use SORT
VARIABLES. 
After doing something like AUTORECODE on the string variables, I need to put
the variables back in the original order.
I tried creating a custom attribute. With something like Pos01 to Pos93 or
#01 to #93 or 1 to 93 as the values. That way I could sort on the custom
variable.

VARIABLE ATTRIBUTE
    VARIABLES =  ALL
    ATTRIBUTE = Input.Order('Pos') arrayname[93].

Variable Attribute No value was specified for attribute "arrayname".

With only 93 variables it is probably faster to type the contents in,
However, I thought others might need something like this, so I am posting.



-----
Art Kendall
Social Research Consultants
--
Sent from: http://spssx-discussion.1045642.n5.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

[Art Kendall]

Thank you.

This will be particularly helpful now that I realize the usefulness of
SORT VARIABLES in cleaning up LABEL VALUES & MISSING attributes.



-----
Art Kendall
Social Research Consultants
--
Sent from: http://spssx-discussion.1045642.n5.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