| Re: multiple delimiters [message #21756 is a reply to message #21677] |
Fri, 15 September 2000 00:00   |
Martin Schultz
Messages: 515
Registered: August 1997
|
Senior Member |
|
|
nrh@imag.wsahs.nsw.gov.au wrote:
>
> Well, its actually a whole heap of strings, most fields separated by
> blanks, and some fields, where there is more than one word, are
> encased by quotation marks. The fields inside the quotes have spaces as
> well, but we want them to be all one field, if you know what I mean.
> Right now we pull out the strings within the quotes, replace all the
> spaces with '_', put it back in, remove the quotes and then we can use
> the strsplit function to remove the extra white spaces created by
> replacing the quotes.
> so, in a nutshell, we have:
> ....PROJECTION-R OTYP DP EXTN img PROC "CM CARDIAC MIBI" .....
> and make it to be(through many painful string ops - it is a huge
> database file)
> PROJECTION-R OTYP DP EXTN img PROC CM_CARDIAC_MIBI ........
> and then we have to arrange it in a struct as every second field is the
> info we actually need. Odd fields are the descriptors.
> Clear as mud?
Your life could be a LOT easier if you had a formatted output, i.e. if
all columns are aligned (I am sure the database that you are using
should be able to produce this). Then you could simply use a formatted
read statement
readf, lun, proj, otyp, dp, extn, img, proc, label, ...,
format='(i4,i6,...,A20,...)'
or (even more elegantly) read into a structure
temp = { proj:1.0, otyp:0L, ..., label:'', ... }
readf, lun, temp, format='...'
It probably boils down to workflow optimization: If you have to do it
once, don't bother and just use a simple code. If you have to do it
several times - always with the same data set - convert the data set
once into a better format (something binary to speed up reading,
optimally a scientific data format for they allow better usability and
are self-describing). If you need to do this several times with
changing data sets, make sure the original data set producer change
their format ;-)
Cheers,
Martin
--
[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[ [[[[[[[
[[ Dr. Martin Schultz Max-Planck-Institut fuer Meteorologie [[
[[ Bundesstr. 55, 20146 Hamburg [[
[[ phone: +49 40 41173-308 [[
[[ fax: +49 40 41173-298 [[
[[ martin.schultz@dkrz.de [[
[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[ [[[[[[[
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
