IF statements in SPSS macros
Locked
 
|
I read that conditional processing in SPSS macros takes this form:
!IF (expression) !THEN statements !ELSE other statements !IFEND However when I try to apply this to my datasets - I can't make it work. I am processing a large number of data sets. They all have the same variable fields but for one variable in consideration for some data sets that variable field will not be populated. When this variable is populated I want the macro to run certain section of code then when the same variable is not populated in another data set the macro will run another section of code. So I have started simple by just testing out my macro using the frequencies command. So I wrote this: DEFINE macro23 () !IF (variable1 = !NULL) !THEN FREQUENCIES VARIABLES=Age /ORDER=ANALYSIS. !ELSE FREQUENCIES VARIABLES=Gender /ORDER=ANALYSIS. !IFEND !ENDDEFINE. macro23 variable1 is the variable that will be blank for some datasets and populated for others. So what my code says is that when variable1 is blank (null) run a frequency on Age otherwise if it is not blank run a frequency on Gender. I notice that when variable1 is system missing I get a frequency on Gender. I expect to get a frequency on age when variable1 is system missing. How can I make it work so I get a frequency on Age when variable1 is not populated. variable1 is a numeric field but I would be interested in seeing what the command would be (if it is any different) if variable1 were to be string. |
Administrator
|
Nope. It doesn't work that way and there is no reason to think it should. Think about it.
!IF (some condition involving some !macro object) !THEN <Do something> !ELSE <do something else> !IFEND DEFINE !blah1 () something !ENDDEFINE. DEFINE !blah2 () somethingelse !ENDDEFINE. DEFINE !Test (arg !TOKENS(1) ) !IF (!arg !EQ 'T') !THEN !blah1 . !ELSE !blah2 . !IFEND !ENDDEFINE. Besides things like FREQUENCIES operate on the entire active case set, you seem to think it would somehow be triggered by a value on a single case. i.e. Macro has absolutely NO clue about contents/values of variables, metadata etc. in fact that damned thing doesn't even know how to add 2 + 2 without having to pull off some silly shit. Of course there is NO way to deliver a return value so it needs to be inline ;-(. DEFINE !add (!pos !TOKENS(1) / !pos !TOKENS(1) ) !LET !sum =!CONCAT(!LENGTH(!BLANKS(!1),!BLANKS(!1)) ECHO !QUOTE(!sum) !ENDDEFINE. !add 2 2 . HTH, DMM
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?" |
|
|
As many know, the Python programmability plugin enables this sort of logic while macro does not. The Programming and Data Management book downloadable from has extensive information on using Python with Statistics (among other topics). On Thu, Mar 9, 2017 at 6:50 AM, David Marso <[hidden email]> wrote: Nope. It doesn't work that way and there is no reason to think it should. -- |
|
Administrator
|
Utterly trivial to employ the 'horrible hack' to aggregate the files and build macro code for insertion contextually WRT the variable state. You would have to do something within python to determine that the variable is empty. Is there a simple function for that?
I was addressing the metaproblem of building procedural logic around individual case states. Really a mixup of levels. --
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?" |
|
|
spssaux2.FindEmptyVars("variablename") From the docstring: Scan specified or all variables and determine which are missing or blank for all cases. Return list of names and optionally delete those variables. So something like (not run) begin program. import spss, spssaux2 if spssaux2.FindEmptyVars('variable1'', alpha=True): spss.Submit("""FREQ Age""") else: spss.Submit("""FREQ Gender""") end program. The alpha parameter says to include string variables in the scan, so if works whether variable1 is numeric or string. The return value from FindEmptyVars will test as true if the variable is empty and otherwise False. On Thu, Mar 9, 2017 at 9:04 AM, David Marso <[hidden email]> wrote: Utterly trivial to employ the 'horrible hack' to aggregate the files and -- |
Re: IF statements in SPSS macros
|
In reply to this post by David Marso
In addition to David's:
Of course, MATRIX-END MATRIX could be effectively used instead of Python to do the OP's task. MATRIX, however, supports variable names (as well as string values) up to 8 bytes (8 characters) length only. And MATRIX has no access to metadata. ===================== 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 |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |