Catastrophic error

[Eric Langston]

I came across an interesting error the other day.  I had one file with a student id, another with student SSN, and another with both.  I needed to get information from the SSN file to the ID file – just doing some basic merging.  I was writing the syntax by hand and it was late on Friday so I was rushing.  I knew that sort cases would execute the pending commands and then run, so I didn’t type EXECUTE between the two commands to save a couple seconds.  Here was my syntax:

 

(file is sorted by ID, but I had forgotten that)

MATCH FILES file=*

/TABLE = ‘dataset2’

/BY ssn.

SORT CASES by ID (a).

 

And SPSS said:

 

The SPSS Statistics Processor is terminating due to a catastrophic error.  Any working data file is lost.  A program error has occurred: A procedure has called OBSERV after having received an end of file indication.  A second pass of the data involves calling OBPINI again.  Please not the circumstances under which this error occurred, attempting to replicate it if possible, and then notify Technical Support.  Restart the processor? (Yes / No)

 

I’ve played around and replicated the problem as follows.  If the match files command returns an error message – duplicate key or file out of order – and then a sort command is issued, then SPSS crashes.  If there is an EXECUTE command between the MATCH FILES and SORT CASES commands, then the error is avoided.

 

Here is syntax that will define a dataset and then generate the error:

 

data list  / var1 1-1 (F) var2 3-3 (F).

begin data

1 8

2 9

4 9

3 8

end data.

 

match files file=* /by var2 /keep=all.

sort cases by var1.

 

Among my friends and colleagues using SPSS, I have one that uses version 12 and another that uses 16.  The syntax did not generate an error for either of them.  I am using SPSS v17.0.1 with all 3 hotfixes installed.  I will be updating to v17.0.2 soon and will see if this has been fixed.

 

Enjoy

 

 

 

---

Eric Langston

Research Analyst

Human Capital Research Corporation

500 Davis Street, Suite 1002

Evanston, IL 60201

[phone]  Phone

[phone]  Fax

 

[Hector Maletta]

I never experienced this in any previous version –from good old SPSS/PC up to v.15 or 16. Thus it is a “feature” introduced with the “upgrade” to v.17, or perhaps it is a “hotfix”. Upgrades, I am learning, are like Eschel stairs: you may be going unknowingly downstairs while you think you are going up.

Hector

 

 

---

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Eric Langston
Sent: 13 April 2009 14:05
To: [hidden email]
Subject: Catastrophic error

 

I came across an interesting error the other day.  I had one file with a student id, another with student SSN, and another with both.  I needed to get information from the SSN file to the ID file – just doing some basic merging.  I was writing the syntax by hand and it was late on Friday so I was rushing.  I knew that sort cases would execute the pending commands and then run, so I didn’t type EXECUTE between the two commands to save a couple seconds.  Here was my syntax:

 

(file is sorted by ID, but I had forgotten that)

MATCH FILES file=*

/TABLE = ‘dataset2’

/BY ssn.

SORT CASES by ID (a).

 

And SPSS said:

 

The SPSS Statistics Processor is terminating due to a catastrophic error.  Any working data file is lost.  A program error has occurred: A procedure has called OBSERV after having received an end of file indication.  A second pass of the data involves calling OBPINI again.  Please not the circumstances under which this error occurred, attempting to replicate it if possible, and then notify Technical Support.  Restart the processor? (Yes / No)

 

I’ve played around and replicated the problem as follows.  If the match files command returns an error message – duplicate key or file out of order – and then a sort command is issued, then SPSS crashes.  If there is an EXECUTE command between the MATCH FILES and SORT CASES commands, then the error is avoided.

 

Here is syntax that will define a dataset and then generate the error:

 

data list  / var1 1-1 (F) var2 3-3 (F).

begin data

1 8

2 9

4 9

3 8

end data.

 

match files file=* /by var2 /keep=all.

sort cases by var1.

 

Among my friends and colleagues using SPSS, I have one that uses version 12 and another that uses 16.  The syntax did not generate an error for either of them.  I am using SPSS v17.0.1 with all 3 hotfixes installed.  I will be updating to v17.0.2 soon and will see if this has been fixed.

 

Enjoy

 

 

 

---

Eric Langston

Research Analyst

Human Capital Research Corporation

500 Davis Street, Suite 1002

Evanston, IL 60201

[phone]  Phone

[phone]  Fax

 

[Matheson, David]

Hello Eric,

  We replicated the problem here with 17.0.1 and 17.0.2 and have filed it as a bug with SPSS Development. We have also replicated the work-around effect of adding the EXECUTE command after the Match Files command.

 

David Matheson

SPSS Technical Support.

 

---

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Eric Langston
Sent: Monday, April 13, 2009 12:05 PM
To: [hidden email]
Subject: Catastrophic error

 

I came across an interesting error the other day.  I had one file with a student id, another with student SSN, and another with both.  I needed to get information from the SSN file to the ID file – just doing some basic merging.  I was writing the syntax by hand and it was late on Friday so I was rushing.  I knew that sort cases would execute the pending commands and then run, so I didn’t type EXECUTE between the two commands to save a couple seconds.  Here was my syntax:

 

(file is sorted by ID, but I had forgotten that)

MATCH FILES file=*

/TABLE = ‘dataset2’

/BY ssn.

SORT CASES by ID (a).

 

And SPSS said:

 

The SPSS Statistics Processor is terminating due to a catastrophic error.  Any working data file is lost.  A program error has occurred: A procedure has called OBSERV after having received an end of file indication.  A second pass of the data involves calling OBPINI again.  Please not the circumstances under which this error occurred, attempting to replicate it if possible, and then notify Technical Support.  Restart the processor? (Yes / No)

 

I’ve played around and replicated the problem as follows.  If the match files command returns an error message – duplicate key or file out of order – and then a sort command is issued, then SPSS crashes.  If there is an EXECUTE command between the MATCH FILES and SORT CASES commands, then the error is avoided.

 

Here is syntax that will define a dataset and then generate the error:

 

data list  / var1 1-1 (F) var2 3-3 (F).

begin data

1 8

2 9

4 9

3 8

end data.

 

match files file=* /by var2 /keep=all.

sort cases by var1.

 

Among my friends and colleagues using SPSS, I have one that uses version 12 and another that uses 16.  The syntax did not generate an error for either of them.  I am using SPSS v17.0.1 with all 3 hotfixes installed.  I will be updating to v17.0.2 soon and will see if this has been fixed.

 

Enjoy

 

 

 

---

Eric Langston

Research Analyst

Human Capital Research Corporation

500 Davis Street, Suite 1002

Evanston, IL 60201

[phone]  Phone

[phone]  Fax