Re: PSPP-BUG: assertion error from src/data/case.c & measurement level is wrong

[Ben Pfaff]

I haven't had a chance to look at your example yet (thank you for sending it!), but I recently reimplemented DELETE VARIABLES because someone else reported a crash related to it. Therefore, it's possible this is already fixed (I'll figure that out quickly when I try it, if so).

On Tue, Feb 28, 2023 at 5:25 PM <knassen@chartermi.net> wrote:

Hi Ben,

 

This one was a bit weird.  I’ll attach the sps and raw data files here and see if you see the issue.

 

I found that it had to do with the order of the STRING command and the DELETE VARIABLES command.  As long as the STRING command is after the DELETE VARIABLES command, it works.   So move STRING GENDERC (A8). after the DELETE VARIABLES command, and it should run.   I’ve had that issue before, but had to move the STRING command to before the DELETE VARIABLES command to get it to run at that time.

 

The data are hierarchical, set up as 5 cards, 59 characters wide.  I’m not using several cards, and those account for the junk1-4 variables that get deleted (that is data from another part of the study).  All the data are numeric except for GENDERC, which is the numeric GENDER variable recoded into Male/Female.  The STRING command makes GENDERC 8 characters wide so data aren’t truncated.

 

I run it from the command line as “pspp -o emotl_test.list -O width=150 emotl_test.sps”

 

I don’t care much about the Measurement Level because we can use the VARIABLE LEVEL command to set that.  It just didn’t do that before, so it was a surprise to see “Unknown.”

 

Kent

 

From: Ben Pfaff <blp@cs.stanford.edu>
Sent: Tuesday, February 28, 2023 7:39 PM
To: knassen@chartermi.net
Cc: bug-gnu-pspp@gnu.org
Subject: Re: PSPP-BUG: assertion error from src/data/case.c & measurement level is wrong

 

Thanks for the report.

 

For the crash, to be able to figure this out, I'll need to look at at least the .sps file and possibly at least some of the data.

 

Unless you assigned a measurement level to these variables, it's normal to see "Unknown".

 

On Tue, Feb 28, 2023 at 4:07 PM <knassen@chartermi.net> wrote:

(latte 5:45pm) knassen: ~/diss> pspp -o emotl.list.test11 -O width=150 emotl.sps

Assertion failed: (idx < c->proto->n_widths), function case_data_idx, file src/data/case.c, line 277.

******************************************************

You have discovered a bug in PSPP.  Please report this

to bug-gnu-pspp@gnu.org.  Please include this entire

message, *plus* several lines of output just above it.

For the best chance at having the bug fixed, also

include the syntax file that triggered it and a sample

of any data file used for input.

proximate cause:     Assertion Failure/Abort

version:             GNU pspp 1.6.2

host_system:         amd64-unknown-freebsd12.3

build_system:        amd64-unknown-freebsd12.3

locale_dir:          /usr/local/share/locale

compiler version:    12.2.0

 

In addition, the Measurement Level column in Display Dictionary output is all “Unknown” values:

 

+--------+--------+----------------------------------------+-----------------+-----+-----+---------+------------+---

|Name    |Position|                  Label                 |Measurement Level| Role|Width|Alignment|Print Format|Wri

+--------+--------+----------------------------------------+-----------------+-----+-----+---------+------------+---

|SUBJECT |       1|Respondent Number                       |Unknown          |Input|    8|Right    |F3.0        |F3.

|STUDY   |       2|Study Code                              |Unknown          |Input|    8|Right    |F1.0        |F1.

|TRTMENT |       3|Treatment Number                        |Unknown          |Input|    8|Right    |F2.0        |F2.

 

_______________________________________________
Bug-gnu-pspp mailing list
Bug-gnu-pspp@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-gnu-pspp