SPSSX Discussion

Error in LOOP

Classic

List

Threaded

14 messages Options

kosetsky

Error in LOOP

Hi!
Kindly ask you to help me with understanding an error that I have with LOOP
function.

My is LOOP is following:

vector str= dCLICK.1.20120315 to email.
LOOP #I=1 to 101.
if cases=1 and str(#I)="" str(#I)=LAG(str(#I)).
END LOOP IF CHAR.INDEX(str(#I),"@")>0.

And I have this error:

>Error # 4002
>A program error has occurred: An impossible situation has been encountered
in
>transformations. This includes invalid operation codes, impossible
subroutine
>arguments, impossible values returned by subroutines, etc. Please note the
>circumstances under which this error occurred, attempting to replicate it
if
>possible, and then notify Technical Support.
>This is an error from which the application cannot recover.
>The SPSS Statistics run will terminate now.

P.S.
vector str defines set of string variables.
Many thanx in advance!

--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Error-in-LOOP-tp5683615.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

David Marso

Re: Error in LOOP

Administrator

Possibilities:
From the FM:
"A single vector must comprise all numeric variables or all string variables. The string
variables must have the same length."
I doubt that the second condition is satisfied! The first might not apply either?
*LAG cannot be applied to VECTORS!!!

DO REPEAT might be a better solution as none of the above limitations apply?
--
It might be useful to describe *WHAT* you are attempting to do.
not sure what the variable "cases" is or its significance.
Looks like you are attempting to fill variables consisting of " " with the value from the preceding row and terminate search upon encountering any field containing a @ in a given row of data?

COMPUTE #found=0.
DO REPEAT str=dCLICK.1.20120315 to email.
+ IF $CASENUM > 1 AND NOT (#found) AND str=" " str=LAG(str).
+ IF CHAR.INDEX(str(#I),"@")>0 #found=1.
END REPEAT.
Does the above solve the problem?
----
It seems (to me) that the SPSS Processor could provide a more informative error message ;-)

kosetsky wrote

Hi!
Kindly ask you to help me with understanding an error that I have with LOOP
function.

My is LOOP is following:

vector str= dCLICK.1.20120315 to email.
LOOP #I=1 to 101.
if cases=1 and str(#I)="" str(#I)=LAG(str(#I)).
END LOOP IF CHAR.INDEX(str(#I),"@")>0.

And I have this error:

>Error # 4002
>A program error has occurred: An impossible situation has been encountered
in
>transformations. This includes invalid operation codes, impossible
subroutine
>arguments, impossible values returned by subroutines, etc. Please note the
>circumstances under which this error occurred, attempting to replicate it
if
>possible, and then notify Technical Support.
>This is an error from which the application cannot recover.
>The SPSS Statistics run will terminate now.

P.S.
vector str defines set of string variables.
Many thanx in advance!

--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Error-in-LOOP-tp5683615.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

Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"

kosetsky

Re: Error in LOOP

Rich Ulrich

Re: Error in LOOP

In reply to this post by kosetsky

> Date: Thu, 3 May 2012 09:04:19 -0700

> From: [hidden email]
> Subject: Error in LOOP
> To: [hidden email]
>
> Hi!
> Kindly ask you to help me with understanding an error that I have with LOOP
> function.
>
> My is LOOP is following:
>
> vector str= dCLICK.1.20120315 to email.
> LOOP #I=1 to 101.
> if cases=1 and str(#I)="" str(#I)=LAG(str(#I)).
> END LOOP IF CHAR.INDEX(str(#I),"@")>0.
>
>
> And I have this error:
>
> >Error # 4002
> >A program error has occurred: An impossible situation has been encountered
> in
> >transformations. This includes invalid operation codes, impossible
> subroutine
> >arguments, impossible values returned by subroutines, etc. Please note the
> >circumstances under which this error occurred, attempting to replicate it
> if
> >possible, and then notify Technical Support.
> >This is an error from which the application cannot recover.
> >The SPSS Statistics run will terminate now.
>
> P.S.
> vector str defines set of string variables.
> Many thanx in advance!
>
> ....

David Marso

Re: Error in LOOP

Administrator

In reply to this post by kosetsky

+ unnecessary for Interactive (syntax window submission), files used in INSERT, but required for INCLUDE (obsolete in newer versions) but useful to visually demarcate nesting levels.
$CASENUM to protect code from trying to LAG first case.
---
example of nesting:

DO REPEAT v=varlist.
+ DO IF (condition1).
+ LOOP.
+ DO IF (condition2).
+ .....
+ END LOOP IF (condition).
+ END IF.
END REPEAT.
---
vs:
DO REPEAT v=varlist.
DO IF (condition1).
LOOP.
DO IF (condition2).
.....
END LOOP IF (condition).
END IF.
END REPEAT.

kosetsky wrote

Thank you very-very much, David!

Final syntax looks even simpler:
DO REPEAT str=dCLICK.1.20120315 to email.
+ IF $CASENUM > 1 AND str="" AND LAG(cases)=1 str=LAG(str).
END REPEAT.
EXECUTE.

The problem was with LAG + vector.
/*IF CHAR.INDEX(str(#I),"@")>0 #found=1.*/ is not necessary now. I need to stop on "email" variable but number of variables in the set is not constant. So, I didn't know how to limit maximum repetition for LOOP.

I spent whole day trying to understand how to make this operation:) Also tried DO REPEAT but smth went wrong.
I'll be greatfull if you open me a secret what "+" is used for and why do I need to put $CASENUM > 1 condition.

Many thanks!

Rick Oliver-3

Re: Error in LOOP

In reply to this post by Rich Ulrich

The application does give an appropriate error message if you try to mix strings and numerics or string of different lengths in the same vector:

Error # 4263 in column 18. Text: z
The list of existing variables on the VECTOR command includes incompatible
variables. A single vector must be composed entirely of numeric variables or
entirely of string variables of the same length.

So apparently something else is going on.

Rick Oliver
Senior Information Developer
IBM Business Analytics (SPSS)
E-mail: [hidden email]
Phone: 312.893.4922 | T/L: 206-4922

From: Rich Ulrich <[hidden email]>
To: [hidden email]
Date: 05/03/2012 01:29 PM
Subject: Re: Error in LOOP
Sent by: "SPSSX(r) Discussion" <[hidden email]>

"... understanding an error" -

Something happened, which definitely should not
have happened, and SPSS found itself in a state
which should *never* arise. The message asks
that you replicate the error: If it happens every
time, then it will not be effect of another program
(virus, say) that has interfered with the contents
of the SPSS program memory in a unique, one-time
manner.

So, it is pretty-much impossible to say what happens
if you can't run it twice and get the same thing.
Ir it replicates, then it might be possible to track it
down. I would add -- re-boot, and run again, because
re-booting manages to clear up a lot of bad behavior
on Windows.

Especially, it would be good if you can confirm it on
another computer, because that would show the problem
is not due to a corruption of the version of SPSS on your
own computer.

Anyway, if the code always does wrong, then
"Notify Technical Support".

David suggest that, perhaps, SPSS allocated a vector
without making sure that all the variables were string,
and of the same length. If that is the problem, then he
gives a work-around; but SPSS should be informed so
that an appropriate error message would be generated
at that earlier mis-step.

--
Rich Ulrich

> Date: Thu, 3 May 2012 09:04:19 -0700
> From: [hidden email]
> Subject: Error in LOOP
> To: [hidden email]
>
> Hi!
> Kindly ask you to help me with understanding an error that I have with LOOP
> function.
>
> My is LOOP is following:
>
> vector str= dCLICK.1.20120315 to email.
> LOOP #I=1 to 101.
> if cases=1 and str(#I)="" str(#I)=LAG(str(#I)).
> END LOOP IF CHAR.INDEX(str(#I),"@")>0.
>
>
> And I have this error:
>
> >Error # 4002
> >A program error has occurred: An impossible situation has been encountered
> in
> >transformations. This includes invalid operation codes, impossible
> subroutine
> >arguments, impossible values returned by subroutines, etc. Please note the
> >circumstances under which this error occurred, attempting to replicate it
> if
> >possible, and then notify Technical Support.
> >This is an error from which the application cannot recover.
> >The SPSS Statistics run will terminate now.
>
> P.S.
> vector str defines set of string variables.
> Many thanx in advance!
>
> ....

Rich Ulrich

Re: Error in LOOP

In reply to this post by David Marso

So the original problem was that a loop to 101 tried to read
(and set values for) too many variables in a vector that was
shorter than 101.

I wonder if SPSS detected this and quit -- in which case,
I agree that there should be a better message -- or if it
carried out instructions to store values in illegal locations,
until it corrupted itself and finally detected another problem.

- This reminds me of what I have read about "exploits"
of some worms and viruses, which read beyond defined
vectors to steal information that should not be accessible.

--
Rich Ulrich

> Date: Thu, 3 May 2012 11:38:11 -0700
> From: [hidden email]
> Subject: Re: Error in LOOP
> To: [hidden email]
> ...
[solution, explained, and comments on solution
from the original poster]

Rick Oliver-3

Re: Error in LOOP

In reply to this post by kosetsky

The plus sign doesn't really do anything...unless you're running in the old "batch" mode, in which case it's just a way to allow indenting of commands. In the standard "interactive" mode, you can indent commands as much as you want without needing to stick something in the first column.

Rick Oliver
Senior Information Developer
IBM Business Analytics (SPSS)
E-mail: [hidden email]
Phone: 312.893.4922 | T/L: 206-4922

From: kosetsky <[hidden email]>
To: [hidden email]
Date: 05/03/2012 04:10 PM
Subject: Re: Error in LOOP
Sent by: "SPSSX(r) Discussion" <[hidden email]>

Thank you very-very much, David! Final syntax looks even simpler: DO REPEAT str=dCLICK.1.20120315 to email. + IF $CASENUM > 1 AND str="" AND LAG(cases)=1 str=LAG(str). END REPEAT. EXECUTE. The problem was with LAG + vector. /*IF CHAR.INDEX(str(#I),"@")>0 #found=1.*/ is not necessary now. I need to stop on "email" variable but number of variables in the set is not constant. So, I didn't know how to limit maximum repetition for LOOP. I spent whole day trying to understand how to make this operation:) Also tried DO REPEAT but smth went wrong. I'll be greatfull if you open me a secret what "+" is used for and why do I need to put $CASENUM > 1 condition. Many thanks! -- View this message in context:http://spssx-discussion.1045642.n5.nabble.com/Error-in-LOOP-tp5683709p5683924.htmlSent 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

kosetsky

Re: Error in LOOP

In reply to this post by Rich Ulrich

My DO REPEAT syntax now works well. But I can't use it if I define a set of both numeric and string variables.
How can I check variable type before using logical expression in transformation to avoid error message?

PS. Regarding the previous error I've had exactly 101 string variables in the set and they were unified by ALTER TYPE function. So, doubt that this could be a reason of error.

Jon K Peck

Re: Error in LOOP

One way to do this would be to define macros for the string and numeric variables and use separate DO REPEAT loops for each.
Statistics has no built-in way to check for variable type, but there is an extension command SPSSINC SELECT VARIABLES that can build macros based on variable metadata.

Here is an example.
SPSSINC SELECT VARIABLES MACRONAME="!numeric"
/PROPERTIES TYPE=NUMERIC.

desc !numeric.

SPSSINC SELECT VARIABLES MACRONAME="!string"
/PROPERTIES TYPE=STRING.

DO REPEAT z = !numeric.
compute z = z*z.
end repeat.

Other selection criteria can also be used in the extension command such as specific names or patterns in the names. See the dialog that appears after installation as Utilities>Define Variable Macro or run SPSSINC SELECT VARIABLES /HELP for details.

To use this extension command, you need to install the Python Essentials for your version of Statistics and the SPSSINC SELECT VARIABLES command. These are available from the SPSS Community website (www.ibm.com/developerworks/spssdevcentral).
Click on Downloads for SPSS Statistics.
On that page, you will see a link to the Python Essentials for various versions of Statistics.
There is also a link to the Extension Commands collection.

If you are on Win7, be sure to do the installs using Run As Administrator or see the FAQ on the site for solving permissions problems.

HTH,

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
new phone: 720-342-5621

From: kosetsky <[hidden email]>
To: [hidden email]
Date: 05/04/2012 07:52 AM
Subject: Re: [SPSSX-L] Error in LOOP
Sent by: "SPSSX(r) Discussion" <[hidden email]>

My DO REPEAT syntax now works well. But I can't use it if I define a set of both numeric and string variables. How can I check variable type before using logical expression in transformation to avoid error message? PS. Regarding the previous error I've had exactly 101 string variables in the set and they were unified by ALTER TYPE function. So, doubt that this could be a reason of error. -- View this message in context:http://spssx-discussion.1045642.n5.nabble.com/Error-in-LOOP-tp5683709p5685426.htmlSent 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

Jim Marks

Re: Error in LOOP

In reply to this post by kosetsky

DO REPEAT supports multiple definitions:

DO REPEAT
x = stringvar1 stringvar99 stringvar23
/y = numvar1 to numvar15
.

You can then process each list of variables with different commands.

Jim Marks
Sr Market Research Manager
National Market Research
Kaiser Foundation Health Plan of the Mid-Atlantic States, Inc.
2101 E. Jefferson St.
Rockville, MD 20852
Phone: (301) 816-6822
Cell Phone: (301) 456-6164

NOTICE TO RECIPIENT: If you are not the intended recipient of this e-mail, you are prohibited from sharing, copying, or otherwise using or disclosing its contents. If you have received this e-mail in error, please notify the sender immediately by reply e-mail and permanently delete this e-mail and any attachments without reading, forwarding or saving them. Thank you.

From: kosetsky <[hidden email]>
To: [hidden email]
Date: 05/04/2012 09:49 AM
Subject: Re: Error in LOOP
Sent by: "SPSSX(r) Discussion" <[hidden email]>

Rick Oliver-3

Re: Error in LOOP

DO REPEAT requires that each stand-in variables have the same number of implied or explicit values. So your example probably won't work.

Rick Oliver
Senior Information Developer
IBM Business Analytics (SPSS)
E-mail: [hidden email]
Phone: 312.893.4922 | T/L: 206-4922

From: Jim Marks <[hidden email]>
To: [hidden email]
Date: 05/04/2012 10:06 AM
Subject: Re: Error in LOOP
Sent by: "SPSSX(r) Discussion" <[hidden email]>

David Marso

Re: Error in LOOP

Administrator

In reply to this post by Jim Marks

Not so fast!
The lists have to be the same length and they are processed in parallel.
That drastically limits such an approach.
The core issue is being able to *NOT* have to manually type in all the variable names after having to sift through and pull strings from numbers.
JoNoH's EXTENSION doohicky is probably the best way.
---

Jim Marks wrote

DO REPEAT supports multiple definitions:

DO REPEAT
x = stringvar1 stringvar99 stringvar23
/y = numvar1 to numvar15
.

You can then process each list of variables with different commands.

Jim Marks
Sr Market Research Manager
National Market Research
Kaiser Foundation Health Plan of the Mid-Atlantic States, Inc.
2101 E. Jefferson St.
Rockville, MD 20852
Phone: (301) 816-6822
Cell Phone: (301) 456-6164

NOTICE TO RECIPIENT: If you are not the intended recipient of this
e-mail, you are prohibited from sharing, copying, or otherwise using or
disclosing its contents. If you have received this e-mail in error,
please notify the sender immediately by reply e-mail and permanently
delete this e-mail and any attachments without reading, forwarding or
saving them. Thank you.

From: kosetsky <[hidden email]>
To: [hidden email]
Date: 05/04/2012 09:49 AM
Subject: Re: Error in LOOP
Sent by: "SPSSX(r) Discussion" <[hidden email]>

My DO REPEAT syntax now works well. But I can't use it if I define a set
of
both numeric and string variables.
How can I check variable type before using logical expression in
transformation to avoid error message?

PS. Regarding the previous error I've had exactly 101 string variables in
the set and they were unified by ALTER TYPE function. So, doubt that this
could be a reason of error.

--
View this message in context:
http://spssx-discussion.1045642.n5.nabble.com/Error-in-LOOP-tp5683709p5685426.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

Jim Marks

Re: Error in LOOP

In reply to this post by Rick Oliver-3

*oops*
I thought the limitation only applied if I wanted to use both lists in the same expression.

DO REPEAT
x = stringvar1 stringvar99 stringvar23 .
<commands.>
END REPEAT

DOREPEAT
/y = numvar1 to numvar15 .
<commands>.
END REPEAT.

Jim Marks
Sr Market Research Manager
National Market Research
Kaiser Foundation Health Plan of the Mid-Atlantic States, Inc.
2101 E. Jefferson St.
Rockville, MD 20852
Phone: (301) 816-6822
Cell Phone: (301) 456-6164

NOTICE TO RECIPIENT: If you are not the intended recipient of this e-mail, you are prohibited from sharing, copying, or otherwise using or disclosing its contents. If you have received this e-mail in error, please notify the sender immediately by reply e-mail and permanently delete this e-mail and any attachments without reading, forwarding or saving them. Thank you.

From: Rick Oliver <[hidden email]>
To: Jim T Marks/MD/KAIPERM@KAIPERM
Cc: [hidden email]
Date: 05/04/2012 11:12 AM
Subject: Re: Error in LOOP

DO REPEAT supports multiple definitions:

DO REPEAT
x = stringvar1 stringvar99 stringvar23
/y = numvar1 to numvar15
.

You can then process each list of variables with different commands.

Jim Marks
Sr Market Research Manager
National Market Research
Kaiser Foundation Health Plan of the Mid-Atlantic States, Inc.
2101 E. Jefferson St.
Rockville, MD 20852
Phone: (301) 816-6822
Cell Phone: (301) 456-6164

NOTICE TO RECIPIENT: If you are not the intended recipient of this e-mail, you are prohibited from sharing, copying, or otherwise using or disclosing its contents. If you have received this e-mail in error, please notify the sender immediately by reply e-mail and permanently delete this e-mail and any attachments without reading, forwarding or saving them. Thank you.

From: kosetsky <[hidden email]>
To: [hidden email]
Date: 05/04/2012 09:49 AM
Subject: Re: Error in LOOP
Sent by: "SPSSX(r) Discussion" <[hidden email]>