Paul's message brings back unfond memories of the hours I wasted until a
friend with SAS opened the file and saved it in SPSS format.
Two comments:
The ICPSR file ends in 1996, which limits it utility. IMF sells a CD
that has historical and current data.
Also, the IMF CD stores the data in an MS-Access file, and includes a
well-designed program to extract data (for the uninitiated). IMHO,
unless a person's research budget is absolutely zero, it's a more
efficient use of time to buy the CD than it is to massage the ICPSR data
into shape.
Joe Doherty
UCLA
Paul Johnson wrote:
> Benjamin Freeman wrote:
>> I am having some difficulty working with the IMF Government Finance
>> Statistics Data from the ICPSR Study No. 8624. Specifically, I do not
>> know
>> exactly how to unpack the main data file. Has anyone written code to
>> unpack
>> this file? Or, does anyone know of an alternative source for this
>> data that
>> does not require unpacking?
>>
> You need to be more specific with your question. I don't have any
> trouble with "unpacking".
>
> $ unzip 6279173.zip
> Archive: 6279173.zip
> extracting: 6279173/ICPSR_08624/08624-manifest.txt
> extracting: 6279173/ICPSR_08624/08624-related_literature.txt
> extracting: 6279173/ICPSR_08624/08624-descriptioncitation.pdf
> inflating: 6279173/ICPSR_08624/DS0001/08624-0001-Codebook.pdf
> inflating: 6279173/ICPSR_08624/DS0001/08624-0001-Data.txt
> inflating: 6279173/ICPSR_08624/DS0002/08624-0002-Data.txt
>
> Now, concerning the file 6279173/ICPSR_08624/DS0001/08624-0001-Data.txt,
> I do see trouble. It is a block of characters I don't recognize. It is
> certainly not encoded in ASCII or Unicode. My colleagues here suspect
> it might be encoded EBCDIC character set, but my first guess was that it
> was edited by somebody in a program that was popular in 1998, say "Word
> Perfect" or such, and that person who saved it forgot to save into plain
> text format. I don't think it is EBCDIC, because:
>
> $ recode EBCDIC..MSDOS test.txt
> recode: test.txt failed: Invalid input in step `ANSI_X3.4-1968..ISO-8859-1'
>
> Ah, then back to read the codebook. I see this comment "The data are
> stored in packed zoned decimal format. A supplemental COBOL processing
> program is available for use with this dataset." The COBOL program
> called Funpack is provided in the directory DS0002, and my guess is that
> one needs a COBOL compiler to run this program. Here are the first few
> lines, in case you need to be convinced.
>
> EBCDIC LINE
> 000010 IDENTIFICATION DIVISION.
> 000020 PROGRAM-ID. 'FUNPACK'.
> 000030 AUTHOR. KATHLEEN NELICK.
> 000040 INSTALLATION. IMF BUREAU OF STATISTICS.
> 000050 DATE-WRITTEN. FEBRUARY 1972.
> 000060 DATE-COMPILED.
> 000070*REMARKS. REFORMAT PACKED IFS DF RECORD TO INTERNAL DF RECORD.
>
> I tracked down a couple of COBOL compilers, but could not make any
> progress compiling that code.
>
> As I google about in the internet, I gather that several IMF datasets
> are in this format and several people have asked what to do about them.
> I don't have more time to spend on this, but if I were you the first
> thing I would try is PROC DATASOURCE in SAS, which claims it can handle
> IMF files.
>
> Good luck. I think if you are more clear about what goes wrong, you
> are probably more likely to get useful answers. When you leave us to
> track down what dataset you are using and then guess what might be going
> wrong, you are asking an awful lot.
>
> --
> Paul E. Johnson email: [log in to unmask]
> Dept. of Political Science http://pj.freefaculty.org
> 1541 Lilac Lane, Rm 504
> University of Kansas Office: (785) 864-9086
> Lawrence, Kansas 66044-3177 FAX: (785) 864-5700
>
> **********************************************************
> Political Methodology E-Mail List
> Editor: Karen Long Jusko <[log in to unmask]>
> **********************************************************
> Send messages to [log in to unmask]
> To join the list, cancel your subscription, or modify
> your subscription settings visit:
>
> http://polmeth.wustl.edu/polmeth.php
>
> **********************************************************
**********************************************************
Political Methodology E-Mail List
Editor: Karen Long Jusko <[log in to unmask]>
**********************************************************
Send messages to [log in to unmask]
To join the list, cancel your subscription, or modify
your subscription settings visit:
http://polmeth.wustl.edu/polmeth.php
**********************************************************
|