describe matrix like data

[Moon Kid]

I am totaly unaware of the needed word for this problem. So I don't
know what to ask for exactly.

I "just" want to describe data like this.

There are a couple of task done in three different time periodes.
It is represented in the data structure like this:
Task-a/PhaseA: vAa
Task-a/PhaseB: vBa
Task-a/PhaseC: vCa
Task-c/PhaseB: vBc
...

I want to display it like this in only one table.

        PhaseA  PhaseB  PhaseC
Task-a    0       3        1
Task-b   .....
Task-c
...

In data it could be described like this.
        PhaseA    PhaseB     PhaseC
Task-a MEAN(vAa)  MEAN(vBa)  MEAN(vCa)
Task-b MEAN(vAb)  MEAN(vBb)  MEAN(vCb)
Task-c MEAN(vAc)  MEAN(vBc)  MEAN(vCc)
...

But I don't know how. I think CTABLES can not handle this.

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

[Andy W]

I don't have CTABLES, but it sounds to me if you have a set of dummy variables representing Task & Phase and then have a numeric value of v?? you can either generate the necessary aggregation or use CTABLES with the appropriate nesting structure.

Something like:

CTABLES
  /TABLE Task BY v?? [MEAN] > Phase.

Here is an example using AGGREGATE to make the table in a physical SPSS dataset (and this shows the necessary dummy variable structure I was talking about).

**************************.
INPUT PROGRAM.
LOOP Id = 1 TO 1000.
END CASE.
END LOOP.
END FILE.
END INPUT PROGRAM.
STRING Task Phase (A20).
COMPUTE Task = CONCAT("Task-",CHAR.SUBSTR("abc",TRUNC(RV.UNIFORM(1,4)),1)).
COMPUTE Phase = CONCAT("Phase",CHAR.SUBSTR("ABC",TRUNC(RV.UNIFORM(1,4)),1)).
COMPUTE v = RV.NORMAL(0,1).
EXECUTE.

*Example CTABLES command (my best guess).
CTABLES
  /TABLE Task BY v [MEAN] > Phase.

*Using aggregate to get the table.
DATASET DECLARE TabMeans.
AGGREGATE OUTFILE = 'TabMeans'
  /BREAK = Task Phase
  /v = MEAN(v).
DATASET ACTIVATE TabMeans.
CASESTOVARS /ID = Task /INDEX Phase.
**************************.

[David Marso]

A specific numeric example might be useful to communicate the actual issue.
In lieu of this I direct you to study the AGGREGATE and CASESTOVARS commands!

Moon Kid wrote

I am totaly unaware of the needed word for this problem. So I don't
know what to ask for exactly.

I "just" want to describe data like this.

There are a couple of task done in three different time periodes.
It is represented in the data structure like this:
Task-a/PhaseA: vAa
Task-a/PhaseB: vBa
Task-a/PhaseC: vCa
Task-c/PhaseB: vBc
...

I want to display it like this in only one table.

        PhaseA  PhaseB  PhaseC
Task-a    0       3        1
Task-b   .....
Task-c
...

In data it could be described like this.
        PhaseA    PhaseB     PhaseC
Task-a MEAN(vAa)  MEAN(vBa)  MEAN(vCa)
Task-b MEAN(vAb)  MEAN(vBb)  MEAN(vCb)
Task-c MEAN(vAc)  MEAN(vBc)  MEAN(vCc)
...

But I don't know how. I think CTABLES can not handle this.

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

[Richard Ristow]

At 12:36 PM 4/22/2014, Moon Kid wrote:

>There are a couple of task done in three different time periodes.
>It is represented in the data structure like this:
>Task-a/PhaseA: vAa
>Task-a/PhaseB: vBa
>Task-a/PhaseC: vCa
>Task-c/PhaseB: vBc
>...
>
>I want to display it like this in only one table.
>In data it could be described like this.
>         PhaseA    PhaseB     PhaseC
>Task-a MEAN(vAa)  MEAN(vBa)  MEAN(vCa)
>Task-b MEAN(vAb)  MEAN(vBb)  MEAN(vCb)
>Task-c MEAN(vAc)  MEAN(vBc)  MEAN(vCc)

This is similar (NOT identical) to a solution I just posted(*). David
is right that it's AGGREGATE followed by CASESTOVARS. Assuming you
start with a file with three variables, namely 'Task', 'Phase', and
'Value', something like this (untested)

DATASET   DECLARE Summary.
AGGREGATE OUTFILE=Summary
    /BREAK= Task Phase
    /MeanVal = MEAN(Value).

CASESTOVARS
    /ID      = Task
    /INDEX   = Phase.


(*) See
Date:    Wed, 23 Apr 2014 17:48:00 -0400
From:    Richard Ristow <[hidden email]>
Subject: Re: Dsiplaying mulitple values
To:      [hidden email]

=====================
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]

Slip
DATASET ACTIVATE summary.
in between the AGG and C2V.

Richard Ristow wrote

At 12:36 PM 4/22/2014, Moon Kid wrote:

>There are a couple of task done in three different time periodes.
>It is represented in the data structure like this:
>Task-a/PhaseA: vAa
>Task-a/PhaseB: vBa
>Task-a/PhaseC: vCa
>Task-c/PhaseB: vBc
>...
>
>I want to display it like this in only one table.
>In data it could be described like this.
>         PhaseA    PhaseB     PhaseC
>Task-a MEAN(vAa)  MEAN(vBa)  MEAN(vCa)
>Task-b MEAN(vAb)  MEAN(vBb)  MEAN(vCb)
>Task-c MEAN(vAc)  MEAN(vBc)  MEAN(vCc)

This is similar (NOT identical) to a solution I just posted(*). David
is right that it's AGGREGATE followed by CASESTOVARS. Assuming you
start with a file with three variables, namely 'Task', 'Phase', and
'Value', something like this (untested)

DATASET   DECLARE Summary.
AGGREGATE OUTFILE=Summary
    /BREAK= Task Phase
    /MeanVal = MEAN(Value).

CASESTOVARS
    /ID      = Task
    /INDEX   = Phase.


(*) See
Date:    Wed, 23 Apr 2014 17:48:00 -0400
From:    Richard Ristow <[hidden email]>
Subject: Re: Dsiplaying mulitple values
To:      [hidden email]

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

[Richard Ristow]

At 07:39 PM 4/23/2014, David Marso wrote:

>Slip
>DATASET ACTIVATE summary.
>in between the AGG and C2V.

Thx.

=====================
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]

I've been bitten before by forgetting that non * files do not have 'focus' after AGGREGATE.

Richard Ristow wrote

At 07:39 PM 4/23/2014, David Marso wrote:

>Slip
>DATASET ACTIVATE summary.
>in between the AGG and C2V.

Thx.

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