| Logic problem [message #58215] |
Wed, 16 January 2008 13:42 |
Ryan.
Messages: 77
Registered: March 2006
|
Member |
|
|
Hi All,
I need assistance with the logic of a particular routine to retrieve a
list of files by date. I have tried a few methods but they haven't been
successful. I'm posting here to hopefully get some assistance on the
logic that I'm using. I can't seem to get it correct. Here are the
details:
I have a directory of folders (As you can see the folder names
correspond to a time period that overlap):
2004-02-19_2004-02-20
2004-02-20_2004-02-21
2004-02-21_2004-02-25
2004-02-25_2004-02-28
2004-03-06_2004-03-10
2004-03-10_2004-03-13
...
Each of these folders contains various files, I am interested in
obtaining one with a particular extension, say *.SAS (I should note,
that sometimes this file does not exist within the folder).
I want to have a special routine that given a date (or a start and end
date) to return the full path of the *.SAS file(s) or if it doesn't
exist to print out a statement saying it doesn't exist. If the desired
dates are spread over 2 or more folders I want it to return all the
paths of the files. If the date desired lands on the overlapping part
of the date (e.g. 2004-02-20 in sample folders above), I want it to
return the path of the two files.
Here are some examples of what I would like returned (using the list of
folders above):
IDL> print, findsasfiles( JULDAY(2,19,2004) )
full_path/2004-02-19_2004-02-20/file.SAS
IDL> print, findsasfiles( JULDAY(2,20,2004) )
full_path/2004-02-19_2004-02-20/file.SAS
full_path/2004-02-20_2004-02-21/file.SAS
IDL> print, findsasfiles( JULDAY(2,19,2004), JULDAY(2,20,2004) )
full_path/2004-02-19_2004-02-20/file.SAS
full_path/2004-02-20_2004-02-21/file.SAS
IDL> print, findsasfiles( JULDAY(2,20,2004), JULDAY(2,25,2004) )
full_path/2004-02-19_2004-02-20/file.SAS
full_path/2004-02-20_2004-02-21/file.SAS
full_path/2004-02-21_2004-02-25/file.SAS
full_path/2004-02-25_2004-02-28/file.SAS
Here is the function so far:
FUNCTION findsasfiles, date, enddate, MISSING=missing, $
NMISSING=nmissing
;Get directory names:
;directory of SAS files
sasdir = FILEPATH('', ROOT_DIR=rch_getrootdir(),
SUBDIRECTORY=['plan'])
;Getting the list of folder names in 'sasdir' directory
dirlist = FILE_SEARCH(sasdir+'*', COUNT=nDirs, /FULLY_QUALIFY_PATH, $
/TEST_DIRECTORY)
;Remove erroneous directory names:
;make sure no extra folder names are found except for the one
;corresponding to a time span
p = STRPOS(dirlist, '200')
diridx = WHERE(p+1, ndirs)
IF ndirs GT 0 THEN dirlist = STRMID(dirlist[WHERE(p+1)], 1#p) $
ELSE dirlist = STRMID(dirlist, 1#p)
;some needed constants
dirs = STRARR(nDirs)
dirdates = DBLARR(nDirs, 2)
;Extract dates from folder names:
FOR i = 0L, ndirs-1 DO BEGIN
splitdir = STRSPLIT(dirlist[i], PATH_SEP(), COUNT=slashcnt)
;retrieve directory name
dirs[i] = STRMID(dirlist[i], splitdir[slashcnt-1])
;retrieve folder name and dates
datesplit = STRSPLIT(dirs[i], '_', /EXTRACT)
dirdates[i,0] = JULDAY(STRMID(datesplit[0],5,2),
STRMID(datesplit[0],8,2), $
STRMID(datesplit[0],0,4))
dirdates[i,1] = JULDAY(STRMID(datesplit[1],5,2),
STRMID(datesplit[1],8,2), $
STRMID(datesplit[1],0,4))
ENDFOR
;******************************** NEED HELP AFTER THIS POINT *****
;Find dates that are searched for:
idx = WHERE(dirdates[*,0] GE date, startcnt)
IF startcnt LT 1 THEN BEGIN
PRINT, 'Start Date Not Found. Returning...'
RETURN, -1S
ENDIF
IF endcorrect THEN BEGIN
endidx = WHERE(dirdates[*,1] GT enddate, endcnt)
IF (endcnt GT 0) THEN BEGIN
idx = [idx[0], endidx[0]]
folders = dirlist[idx[0]:idx[1]]
ENDIF
ENDIF ELSE folders = dirlist[idx[0]]
;Discover if SAS file exists and return it if it does:
nFiles = FIX(N_ELEMENTS(folders))
files = STRARR(nFiles)
missing = STRARR(nFiles)
sascounter = 0S
nMissing = 0S
FOR j=0L, nFiles-1 DO BEGIN
sasfind = FILE_SEARCH(sasdir+folders[j], '*.SAS',
COUNT=sasfindcount, /FULLY_QUALIFY_PATH)
IF (sasfindcount GT 0) THEN BEGIN
files[sascounter] = sasfind
sascounter += 1
ENDIF ELSE BEGIN
PRINT, 'No SAS file found in folder: '+folders[j]
missing[nMissing] = folders[j]
nMissing += 1
ENDELSE
ENDFOR
CASE 1 OF
(nMissing EQ 0) AND (sascounter EQ 0): BEGIN
files = -1S
missing = -1S
END
(nMissing GT 0) AND (sascounter EQ 0): BEGIN
files = -1S
print, 'No files found'
END
(nMissing EQ 0) AND (sascounter GT 0): BEGIN
missing = -1S
print, 'All files found'
END
ELSE: BEGIN
files = files[0:(sascounter-1)]
missing = missing[0:(nMissing-1)]
PRINT, 'A bit of this, a bit of that'
END
ENDCASE
;******************
RETURN, files
END
------------------
The routine as it stands doesn't quite work because the indices returned
(the final 'idx' variable) can be backwards (eg. [175,174]) and thus
cause an error when trying to execute the line "folders =
dirlist[idx[0]:idx[1]]".
To me it doesn't seem like an overly difficult problem but I've spent
the last 2 or 3 days trying to get it right with no success. I need
some new minds to help me with this.
Thanks,
Ryan.
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
