Date:         Thu, 2 Dec 2004 12:35:27 +0200
Reply-To:     Martins.Liberts@csb.gov.lv
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Martins Liberts <Martins.Liberts@csb.gov.lv>
Subject:      Re: "/" after the macro calls
Content-Type: text/plain; charset="us-ascii"

I just found out that the trick with slash (/) doesn't work in AGGREGATE.

Example1. It is OK.

agg out=!dataq+"\Work\out_"+!varq+"_"+!subgrq+".sav"
 /break !subgr /outl=sum(!var_nr)
 /!temp /outl2=sum(!var_nr).

Example2. Should work, but it doesn't.

agg out=!dataq+"\Work\out_"+!varq+"_"+!subgrq+".sav"
 /break !subgr/
 /outl=sum(!var_nr)
 /!temp/
 /outl2=sum(!var_nr).

2396 M>
agg out=!dataq+"\Work\out_"+!varq+"_"+!subgrq+".sav"
2397 M>  agg out='C:\My Documents\Darbs\IntraStat\data_200407\Reize2\Data'
2398 M>   +"\Work\out_"+'EXP'
2399 M>   +"_"+'NOD'
 /break !subgr/
2400 M>   +".sav"
2401 M>    /break NOD
2402 M>   /
 /outl=sum(!var_nr)
2403 M>    /outl=sum(EXP_67
2404 M>   )
 /!temp/
2405 M>    /N=N
2406 M>   /
 /outl2=sum(!var_nr).
2407 M>    /outl2=sum(EXP_67
2408 M>   ).

>Error # 10917 in column 3.  Text: /
>SPSS expected the definition of a new variable on the AGGREGATE command
but
>did not find a new variable name.
>This command not executed.

>Error # 10917 in column 3.  Text: /
>SPSS expected the definition of a new variable on the AGGREGATE command
but
>did not find a new variable name.

It is strange because in MATCH FILES command the line with "/" doesn't
couse an error.

2169 M>
match files
2170 M>  match files
 /file=!pathq+"\data64\Ready\"+!varq+"_"+!subgrq+".sav"
2171 M>    /file='C:\My Documents\Darbs\IntraStat'
2172 M>   +"\data64\Ready\"+'EXP'
2173 M>   +"_"+'NOD'
 /in prev
2174 M>   +".sav"
2175 M>    /in prev
 /file=*
2176 M>    /file=*
 /rename !var=!var_nr/
2177 M>    /rename EXP
2178 M>   =EXP_67
2179 M>   /
 /by nmk !subgr/
2180 M>    /by nmk NOD
2181 M>   /
 /keep nmk !subgr !varl !var_nr.
2182 M>    /keep nmk NOD
2183 M>    EXP_59 EXP_60 EXP_61 EXP_62 EXP_63 EXP_64
2184 M>    EXP_67
2185 M>   .

Probably it is because in AGGREGATE after the /break subcommand there
could be only subcommands for definitions of a new variables. In MATCH
FILES after /rename and /by subcommands (as in the example) there could be
different subcommands.

--
Martins Liberts
Central Statistical Bureau of Latvia
address: Lacplesa iela 1, Riga, LV-1301, Latvia
phone: 371-7366877
email: Martins.Liberts@csb.gov.lv




Richard Ristow <wrristow@mindspring.com>
Sent by: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
01.12.2004 19:51
Please respond to Richard Ristow


        To:     SPSSX-L@LISTSERV.UGA.EDU
        cc:
        Subject:        Re: Macros and importing data   ["/" after the macro calls]


At 11:44 AM 12/1/2004, David Hutton asked:

>Richard:
>You wrote:
>_______________________________________________________________________
>>As for the macros inserting newlines, yes, they do. I don't know why
>>it
>>was done that way, but it's one of the things I like least about SPSS
>>macros. It means that the following works:
>>
>>MATCH FILES
>>     /FILE=   !D_SAVE(HV_Base)/
>>     /FILE=   !D_SAVE(IntvSmry)/
>>     /BY   FAMKEY HV_DT VISIT# ACC_DB.
>>
>>and the following doesn't:
>>
>>MATCH FILES
>>     /FILE=   !D_SAVE(HV_Base)
>>     /FILE=   !D_SAVE(IntvSmry)
>>     /BY   FAMKEY HV_DT VISIT# ACC_DB.
>>
>>The only difference is leaving out "/" after the macro calls.
>_______________________________________________________________________
>
>Would you mind expanding on this, please?
>Does !D_SAVE(HV_Base) expand to something like
>'C:\temp\myfile.sav'   ?

Yes, exactly.

>Can I assume that _any_  multi-line command would be affected?

And again, yes. If there's interpreted text following the macro call,
on the same line, that text is interpreted as if it were on a
continuation of the command line (meaning, it's handled normally). If
there's no interpreted text (a comment is NOT good enough), a blank
line is effectively inserted, and later lines meant as part of the
command are considered "unrecognized keywords".

>What if the command (e.g. MATCH FILES) is itself within a macro?

If there's a macro VARIABLE as part of a command within a macro, no
extra new-line is inserted. If a macro INVOCATION is part of a command
within a macro, I'm not positive, but I'd count, myself, on its
inserting the newline.

I have no idea why macro expansions do insert newlines. I learned it
the hard way, myself. I've never found it described in documentation,
or in notes from SPSS people. I don't know whether it's bug or feature.
I've assumed, feature; that is, they meant it that way, for some
reason. But, if so, I don't know the reason.

-Best regards,
  Richard