Simple (I think) problem with a Macro

[David Futrell]

Hi everyone:

I have an annoying problem with a macro that I hope someone can help me solve.
The following syntax works perfectly.

DEFINE !REPORTMACRO().

!DO !I=1 !TO 865
temporary.
select if any(!I,g4id1 to g4id9).
compute report_owner=!I.
save outfile=!QUOTE(!CONCAT("D:\Monthly Pulse Survey\G4 Reports\ID ",!I,".sav."))
/keep report_owner empid.
!DOEND

!ENDDEFINE

!REPORTMACRO. 

However, I want to use a calculated value in the !DO line instead of the constant (in this case 865) because this value varies across reporting periods and I currently have to calculate it and manually enter the value into the macro.

I can easily calculate what the end value should be using the aggregate command. Here's what I've tried to do:


compute max_x=max(g4id1 to g4id10).
AGGREGATE
  /OUTFILE=* MODE=ADDVARIABLES
  /BREAK=
  /G4max=MAX(max_x).


DEFINE !REPORTMACRO().

!DO !I=1 !TO G4max
temporary.
select if any(!I,g4id1 to g4id9).
compute report_owner=!I.
save outfile=!QUOTE(!CONCAT("D:\Monthly Pulse Survey\G4 Reports\ID ",!I,".sav."))
/keep report_owner empid.
!DOEND

!ENDDEFINE


!REPORTMACRO. 

The calculated value (G4max) is exactly 865, but when I try the syntax, I get this message:


>Warning # 228
>Invalid string found where numeric value required in macro expression.
 
>Warning # 221
>Macro expansion found an error interpreting an operation code.  Check the
>arguments to functions and loop parameters for valid values.
 
>Error # 1.  Command name: !ERROR_MACRO
>The first word in the line is not recognized as an SPSS Statistics command.
>This command not executed.

Any ideas?

David Futrell
Eli Lilly and Company

[Albert-Jan Roskam]

Hi!

You're trying to assigning to the string "G4max", which has no meaning. It does not refer to the maximum. You could use WRITE and INSERT to assign it to a macro, but that's extremely ugly, so here's an alternative solution:

file handle f / name = 'c:/program files/spss/employee data.sav'.
get file = f.
compute dummy = 1.
aggr out = * / break = dummy / max = max (educ).
add files / file = * / keep = max.

begin program.
import spss
cur = spss.Cursor()
max = int(cur.fetchone()[0])
cur.close()

for period in range(1, max+1):
  spss.Submit("get file = f.")
  #spss.Submit("numeric g4id4 to g4id9 (f4).") # not needed in your dataset, only for testing
  spss.Submit("temporary.")
  spss.Submit("select if any(%s, g4id4 to g4id9)." % period)
  spss.Submit("compute report_owner = %s." % period)
  spss.Submit("save outfile = 'd:/temp/id_%02d.sav'." % period)
end program.

Cheers!!
Albert-Jan

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Before you criticize someone, walk a mile in their shoes, that way
when you do criticize them, you're a mile away and you have their shoes!
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


--- On Thu, 10/8/09, David Futrell <[hidden email]> wrote:

> From: David Futrell <[hidden email]>
> Subject: [SPSSX-L] Simple (I think) problem with a Macro
> To: [hidden email]
> Date: Thursday, October 8, 2009, 4:25 PM
>
> Hi everyone:
>
> I have an annoying problem
> with a macro that I hope someone can help me solve.
>
> The following syntax works
> perfectly.
>
> DEFINE !REPORTMACRO().
>
> !DO !I=1 !TO 865
> temporary.
> select if any(!I,g4id1 to g4id9).
>
> compute report_owner=!I.
> save
>  outfile=!QUOTE(!CONCAT("D:\Monthly Pulse
> Survey\G4 Reports\ID
> ",!I,".sav."))
> /keep report_owner empid.
> !DOEND
>
> !ENDDEFINE
>
> !REPORTMACRO.
>
>
>
> However, I want to use a calculated value in the !DO line
> instead of
> the constant (in this case 865) because this value varies
> across reporting periods and I currently have to calculate
> it and manually enter the value into the macro.
>
> I can easily calculate what
> the end value should be using the aggregate command.
> Here's what I've tried to do:
>
>
> compute max_x=max(g4id1 to g4id10).
> AGGREGATE
>   /OUTFILE=* MODE=ADDVARIABLES
>
>  /BREAK=
>
> /G4max=MAX(max_x).
>
>
> DEFINE
> !REPORTMACRO().
>
> !DO !I=1 !TO G4max
> temporary.
> select if any(!I,g4id1 to g4id9).
>
> compute report_owner=!I.
> save
> outfile=!QUOTE(!CONCAT("D:\Monthly Pulse
> Survey\G4 Reports\ID
> ",!I,".sav."))
> /keep report_owner empid.
> !DOEND
>
> !ENDDEFINE
>
>
> !REPORTMACRO.
>
>
> The calculated value (G4max)
> is exactly 865, but when I try the syntax, I get this
> message:
>
>
> >Warning # 228
> >Invalid string found where
> numeric value required in macro expression.
>
> >Warning # 221
> >Macro expansion found an error
> interpreting an operation code.  Check the
> >arguments to functions and loop
> parameters for valid values.
>
> >Error # 1.
>  Command name: !ERROR_MACRO
> >The first word in the line is
> not recognized as an SPSS Statistics command.
> >This command not
> executed.
>
> Any ideas?
>
> David Futrell
> Eli Lilly and
>  Company
>
>
>
>
>
>

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

[mpirritano]

To all listers!

This statement is not working.


This is all on one line:
if (elig_period = 1) & ((fromdatesvc_s >= app_30_days.1) &
(fromdatesvc_s >=
elig_start_date.1) &

This is on a second line indented 31 spaces:
(fromdatesvc_s < app_first_of_month.1) & (fromdatesvc_s <=
elig_end_date.1)) no_pay_30 = 1.

It is supposed to identify a claim that occurred with eligibility period
1 (individuals can have more than one eligibility period). It should
then flag only those claims that are within 30 days of the first of the
month that the application for the program was filed, where the date of
service is greater than the date of the beginning of eligibility, is
less than the first of the month when the application was completed, and
less than the end of the eligibility period.

The problem is that the statement seems to be ignoring some
stipulations. I've been able to identify that it is not selecting dates
of service that are less than the first of the month that the
application took place in. Could my indentation have something to do
with it? I indent all the time to make my code more readable!

Thanks
Matt



Matthew Pirritano, Ph.D.
Research Analyst IV
Medical Services Initiative (MSI)
Orange County Health Care Agency
[phone]

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

[ViAnn Beadle]

You cannot directly transfer a data value to a macro in this way. You are getting this error because G4max in the macro is just a string value. Macros are really, really limited (e.g., dumb). In the past clever SPSSers would do things like generate macros on the fly by using the WRITE command to create a syntax file with some interspersed data values. Now, the best way to do this is to use Python to generate the syntax and run it because Python knows about the contents of the dictionary (variable names and such) and can retrieve case values.

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of David Futrell
Sent: Thursday, October 08, 2009 8:25 AM
To: [hidden email]
Subject: Simple (I think) problem with a Macro

 

Hi everyone:

I have an annoying problem with a macro that I hope someone can help me solve.
The following syntax works perfectly.

DEFINE !REPORTMACRO().

!DO !I=1 !TO 865
temporary.
select if any(!I,g4id1 to g4id9).
compute report_owner=!I.
save outfile=!QUOTE(!CONCAT("D:\Monthly Pulse Survey\G4 Reports\ID ",!I,".sav."))
/keep report_owner empid.
!DOEND

!ENDDEFINE

!REPORTMACRO. 

However, I want to use a calculated value in the !DO line instead of the constant (in this case 865) because this value varies across reporting periods and I currently have to calculate it and manually enter the value into the macro.

I can easily calculate what the end value should be using the aggregate command. Here's what I've tried to do:


compute max_x=max(g4id1 to g4id10).
AGGREGATE
  /OUTFILE=* MODE=ADDVARIABLES
  /BREAK=
  /G4max=MAX(max_x).


DEFINE !REPORTMACRO().

!DO !I=1 !TO G4max
temporary.
select if any(!I,g4id1 to g4id9).
compute report_owner=!I.
save outfile=!QUOTE(!CONCAT("D:\Monthly Pulse Survey\G4 Reports\ID ",!I,".sav."))
/keep report_owner empid.
!DOEND

!ENDDEFINE


!REPORTMACRO. 

The calculated value (G4max) is exactly 865, but when I try the syntax, I get this message:


>Warning # 228
>Invalid string found where numeric value required in macro expression.
 
>Warning # 221
>Macro expansion found an error interpreting an operation code.  Check the
>arguments to functions and loop parameters for valid values.
 
>Error # 1.  Command name: !ERROR_MACRO
>The first word in the line is not recognized as an SPSS Statistics command.
>This command not executed.

Any ideas?

David Futrell
Eli Lilly and Company

 

[Maguin, Eugene]

Matt,

Maybe somebody else can see the problem without seeing sample data but it
would sure be helpful if you could post data where the selection failed and
where it suceeded. And, probably some explanation of the variables. It seems
to me that since you got not error statement, you must not have any error
related to indention (or any other cause). And, in any event, indention as
part of a continuation should not give an error.

Thanks, Gene Maguin


>>This statement is not working.

This is all on one line:
if (elig_period = 1) & ((fromdatesvc_s >= app_30_days.1) &
(fromdatesvc_s >=
elig_start_date.1) &

This is on a second line indented 31 spaces:
(fromdatesvc_s < app_first_of_month.1) & (fromdatesvc_s <=
elig_end_date.1)) no_pay_30 = 1.

It is supposed to identify a claim that occurred with eligibility period
1 (individuals can have more than one eligibility period). It should
then flag only those claims that are within 30 days of the first of the
month that the application for the program was filed, where the date of
service is greater than the date of the beginning of eligibility, is
less than the first of the month when the application was completed, and
less than the end of the eligibility period.

The problem is that the statement seems to be ignoring some
stipulations. I've been able to identify that it is not selecting dates
of service that are less than the first of the month that the
application took place in. Could my indentation have something to do
with it? I indent all the time to make my code more readable!

Thanks
Matt



Matthew Pirritano, Ph.D.
Research Analyst IV
Medical Services Initiative (MSI)
Orange County Health Care Agency
[phone]

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

[Bruce Weaver]

David Futrell wrote

Hi everyone:

I have an annoying problem with a macro that I hope someone can help me solve.
The following syntax works perfectly.

DEFINE !REPORTMACRO().

!DO !I=1 !TO 865
temporary.
select if any(!I,g4id1 to g4id9).
compute report_owner=!I.
save
 outfile=!QUOTE(!CONCAT("D:\Monthly Pulse Survey\G4 Reports\ID ",!I,".sav."))
/keep report_owner empid.
!DOEND

!ENDDEFINE

!REPORTMACRO. 


However, I want to use a calculated value in the !DO line instead of
the constant (in this case 865) because this value varies across reporting periods and I currently have to calculate it and manually enter the value into the macro.

I can easily calculate what the end value should be using the aggregate command. Here's what I've tried to do:


compute max_x=max(g4id1 to g4id10).
AGGREGATE
  /OUTFILE=* MODE=ADDVARIABLES
 
 /BREAK=
  /G4max=MAX(max_x).


DEFINE !REPORTMACRO().

!DO !I=1 !TO G4max
temporary.
select if any(!I,g4id1 to g4id9).
 
compute report_owner=!I.
save outfile=!QUOTE(!CONCAT("D:\Monthly Pulse Survey\G4 Reports\ID ",!I,".sav."))
/keep report_owner empid.
!DOEND

!ENDDEFINE


!REPORTMACRO. 

The calculated value (G4max) is exactly 865, but when I try the syntax, I get this message:


>Warning # 228
>Invalid string found where numeric value required in macro expression.
 
>Warning # 221
>Macro expansion found an error interpreting an operation code.  Check the
>arguments to functions and loop parameters for valid values.
 
>Error # 1. 
 Command name: !ERROR_MACRO
>The first word in the line is not recognized as an SPSS Statistics command.
>This command not executed.

Any ideas?

David Futrell
Eli Lilly and
 Company

Here's a method that seemed to work on an example I tried with the GSS data file.  (I.e., there were no error messages when I ran the lines that define the macro, including the INCLUDE FILE line.  But I couldn't actually run the macro, given that I didn't have the right data file.)  It's probably not as neat as the Python method that was posted...but it does entail only a fairly minor tweak on what you've already got.


compute max_x=max(g4id1 to g4id10).
AGGREGATE
  /OUTFILE=* MODE=ADDVARIABLES
  /BREAK=
  /G4max=MAX(max_x).

* Write the first two lines of the macro definition to a text file .

do if $casenum EQ 1.
- write outfile = "D:\myfolder\macro_lines.txt" /
 "DEFINE !REPORTMACRO()." / "!DO !I=1 !TO ", G4max .
end if.
exe.

* Now use INCLUDE FILE to bring in the first two lines of the macro.

include file = "D:\myfolder\macro_lines.txt" . /* first two lines of macro definition .
* Now the rest of the macro definition .
temporary.
select if any(!I,g4id1 to g4id9).
compute report_owner=!I.
save outfile=!QUOTE(!CONCAT("D:\Monthly Pulse Survey\G4 Reports\ID ",!I,".sav."))
/keep report_owner empid.
!DOEND

!ENDDEFINE

[Bruce Weaver]

Pirritano, Matthew-2 wrote

To all listers!

This statement is not working.


This is all on one line:
if (elig_period = 1) & ((fromdatesvc_s >= app_30_days.1) &
(fromdatesvc_s >=
elig_start_date.1) &

This is on a second line indented 31 spaces:
(fromdatesvc_s < app_first_of_month.1) & (fromdatesvc_s <=
elig_end_date.1)) no_pay_30 = 1.

It is supposed to identify a claim that occurred with eligibility period
1 (individuals can have more than one eligibility period). It should
then flag only those claims that are within 30 days of the first of the
month that the application for the program was filed, where the date of
service is greater than the date of the beginning of eligibility, is
less than the first of the month when the application was completed, and
less than the end of the eligibility period.

The problem is that the statement seems to be ignoring some
stipulations. I've been able to identify that it is not selecting dates
of service that are less than the first of the month that the
application took place in. Could my indentation have something to do
with it? I indent all the time to make my code more readable!

Thanks
Matt

Hi Matt.  I find that for complicated conditional statements, it often helps to make the statement more readable (by reducing clutter) if I use a bunch of scratch variables for the various conditions.  E.g.,

compute #ep1 = (elig_period EQ 1).
compute #fdGEa30 = (fromdatesvc_s GE app_30_days.1).
compute #fdGEesd1 = (fromdatesvc_s GE elig_start_date.1).
etc...
if #ep1 & (#fdGEa30 & #fdGEesd1 & ...

[Bruce Weaver]

ViAnn Beadle wrote

You cannot directly transfer a data value to a macro in this way. You are
getting this error because G4max in the macro is just a string value. Macros
are really, really limited (e.g., dumb). In the past clever SPSSers would do
things like generate macros on the fly by using the WRITE command to create
a syntax file with some interspersed data values. Now, the best way to do
this is to use Python to generate the syntax and run it because Python knows
about the contents of the dictionary (variable names and such) and can
retrieve case values.

This is not meant in any way to question the merits of Python--I have no doubt that it is superior in many ways to the macro language.  However, one clear advantage of macros is that they do not require installation of Python.  This may not be an issue for some folks, but I know that many of the colleagues with whom I share code will never install Python (or anything else) that is not part of the base installation.  It's hard enough getting some of them to install the latest patches.  ;-)

[ViAnn Beadle]

There are an increasingly large number of additional "packaged" python
programs available on SPSS Devcentral that someday your colleagues might
find to be very, very useful. Unfortunately, finding those programs and what
they actually do is a bit complicated. I have to resort to searches on the
text in a very long page of modules and I don't think that Google has any
bots running around on the site to help.

The common problems of open-source initiatives like this seem to be the
short update cycles with version incompatibilities, decentralization of code
(multiple installs, copying, etc.), less than comprehensive or hard to find
documentation, and the decentralization of information on access.

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Bruce Weaver
Sent: Friday, October 09, 2009 5:19 AM
To: [hidden email]
Subject: Re: Simple (I think) problem with a Macro

ViAnn Beadle wrote:

>
> You cannot directly transfer a data value to a macro in this way. You are
> getting this error because G4max in the macro is just a string value.
> Macros
> are really, really limited (e.g., dumb). In the past clever SPSSers would
> do
> things like generate macros on the fly by using the WRITE command to
> create
> a syntax file with some interspersed data values. Now, the best way to do
> this is to use Python to generate the syntax and run it because Python
> knows
> about the contents of the dictionary (variable names and such) and can
> retrieve case values.
>

This is not meant in any way to question the merits of Python--I have no
doubt that it is superior in many ways to the macro language.  However, one
clear advantage of macros is that they do not require installation of
Python.  This may not be an issue for some folks, but I know that many of
the colleagues with whom I share code will never install Python (or anything
else) that is not part of the base installation.  It's hard enough getting
some of them to install the latest patches.  ;-)



-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/
"When all else fails, RTFM."

NOTE:  My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.
--
View this message in context:
http://www.nabble.com/Simple-%28I-think%29-problem-with-a-Macro-tp25805207p2
5819389.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

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

[Bruce Weaver]

ViAnn Beadle wrote

There are an increasingly large number of additional "packaged" python
programs available on SPSS Devcentral that someday your colleagues might
find to be very, very useful. Unfortunately, finding those programs and what
they actually do is a bit complicated. I have to resort to searches on the
text in a very long page of modules and I don't think that Google has any
bots running around on the site to help.

The common problems of open-source initiatives like this seem to be the
short update cycles with version incompatibilities, decentralization of code
(multiple installs, copying, etc.), less than comprehensive or hard to find
documentation, and the decentralization of information on access.

As noted previously, I don't doubt at all the usefulness of Python.  But to reinforce the point I made earlier, here is the result of a quick poll I just took (n=1).  My office neighbour, a long-time SPSS user, has never even heard of Python or R.  And she is by no means a novice point-and-click user.  Rather, she is quite well-versed in the use of syntax for descriptive stats, graphs, regression models (linear, logistic), ANOVA, etc.  She even includes comments in her syntax.  AFAIK, she has never used macros or the matrix language.  In short, I think she is quite representative of a lot of academic users.  I also think it is easy to forget that those of us  who lurk on mailing lists like this (or in the usenet newsgroup) are probably not very representative of most users.

Bruce

[ViAnn Beadle]

It's always easy to pick your sample;-) Here's what she might be missing:

- Data and dictionary aware macros (which started this thread off). Python
can even provide a point and click interface for your macro.
- Casewise transformations not available via standard SPSS transformation
commands and functions
- Database functions not available via MATCH FILES or ADD FILES
- Programmed pivot table edits including pivoting, hiding, merging tables,
and cell properties.
- Specialized statistical procedures

I seems to recall at least 10 people looking for raking on this list and
that's available via a packaged application. Here are some applications
which posters on this list have been looking for:

- Case control matching
- TURF analysis
- Partial Least Squares
- Confidence Intervals for Proportion
- Compare Datasets

And we shouldn't forget about the interface to R.



-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Bruce Weaver
Sent: Friday, October 09, 2009 10:06 AM
To: [hidden email]
Subject: Re: Python (was Simple (I think) problem with a Macro)

ViAnn Beadle wrote:

>
> There are an increasingly large number of additional "packaged" python
> programs available on SPSS Devcentral that someday your colleagues might
> find to be very, very useful. Unfortunately, finding those programs and
> what
> they actually do is a bit complicated. I have to resort to searches on the
> text in a very long page of modules and I don't think that Google has any
> bots running around on the site to help.
>
> The common problems of open-source initiatives like this seem to be the
> short update cycles with version incompatibilities, decentralization of
> code
> (multiple installs, copying, etc.), less than comprehensive or hard to
> find
> documentation, and the decentralization of information on access.
>
>

As noted previously, I don't doubt at all the usefulness of Python.  But to
reinforce the point I made earlier, here is the result of a quick poll I
just took (n=1).  My office neighbour, a long-time SPSS user, has never even
heard of Python or R.  And she is by no means a novice point-and-click user.
Rather, she is quite well-versed in the use of syntax for descriptive stats,
graphs, regression models (linear, logistic), ANOVA, etc.  She even includes
comments in her syntax.  AFAIK, she has never used macros or the matrix
language.  In short, I think she is quite representative of a lot of
academic users.  I also think it is easy to forget that those of us  who
lurk on mailing lists like this (or in the usenet newsgroup) are probably
not very representative of most users.

Bruce


-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/
"When all else fails, RTFM."

NOTE:  My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.
--
View this message in context:
http://www.nabble.com/Simple-%28I-think%29-problem-with-a-Macro-tp25805207p2
5823745.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

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