[R] A read.table mystery (data for Filemaker Mac)

Emmanuel Charpentier charpent at bacbuc.dyndns.org
Fri Oct 12 09:22:54 CEST 2007


Mark Wardle a écrit :
> 1. Which version of Filemaker? NB: Framemaker is a different program
> (desktop publishing), so do be a little precise!

Dunno. The file is named "export.fm7" ; one might be tempted to infer
Filemaker 7.

"Framemaker" is a typo....

> 2. If it is an ancient version, then I suggest exporting a block of
> columns at a time, and then using merge() in R to join it all back up

Not an option : I do not make the export myself, and I do not have
Filemaker on any machine I can lay my hands on...

> 3. I store all my clinical data in Filemaker 8.5 on the Mac. It is
> great. There have been no significant data export or import issues.

Except for date formats (DD/MM/YYYY in lieu of YYYY-MM-DD), numeric
values (comma as decimal mark) and character set (something looking like
Latin-1 in lieu of utf8). <Sigh ...>

> One problem (with this version - may be fixed in 9.0) is that
> Filemaker ODBC drivers are pretty dreadful and so I do not use ODBC to
> access the data held in a Filemaker database. However, Filemaker's
> client ODBC access works fine, and, in conjunction with some
> commercial ODBC drivers for PostgreSQL, export my data into
> postgresql, and then from there import data into R. It is fast and
> works very well, albeit in a rather convoluted fashion!

That would be a (good !) option if both the "data producer" and me had
access to a common database server. It turns out not to be the case
(alas...).

As a general comment, I fpound the combination of a DBMS (Postgres in my
case), an ODBC-able front end (OOo base, MS-Access, Filemaker : pîck
ypouir poison...) and R a very good working setup. I use it every time I
can. However, in this case, that's pipedream...

> Best wishes,
> 
> Mark
> 
> 
> 
> On 12/10/2007, Emmanuel Charpentier <charpent at bacbuc.dyndns.org> wrote:
>> Gabor Grothendieck a écrit :
>>> On 10/11/07, Emmanuel Charpentier <charpent at bacbuc.dyndns.org> wrote:
>>>> I have to read some clinical data a file coming from Filemaker on
>>>> Macintosh (Ugh ! But it could be worse and come from Excel...).
>>>> Exporting via Excel is out of question since the file has 467 columns
>>>> and 121 lines (+ headers), which is out of reach of Excel. So I received
>>>> an "mer"" files, which is what Filemaker exports as a text file.
>>> The limits for Excel are 1 million rows and 16,000 columns (based on
>>> Excel 2007 which is what I have).
>> I don't. More important, the version of Framemaker used by my
>> correspondent still uses old Excel limits (,o more than 256 columns).
>>
>>> Read the last line on every message to r-help,
>> I do ...
>>
>>                                                 Emmanuel Charpentier
>>
>> ______________________________________________
>> R-help at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>>
>>
>> ______________________________________________________________________
>> This email has been scanned by the MessageLabs Email Security System.
>> For more information please visit http://www.messagelabs.com/email
>> ______________________________________________________________________
>>
> 
>



More information about the R-help mailing list