;+
;PROCEDURE: loadcdfstr
;PURPOSE:
; loads data from specified cdf file into a structure.
;INPUT:
; x: A named variable to return the structure in
; novardata: A named variable to return the non-varying
; data.
;
;KEYWORDS:
; FILENAMES: [array of] CDF filename[s]. (or file id's)
; PATH: CDF file path.
; VARNAMES: [array of] CDF variable name[s] to be loaded.
; NOVARNAMES: [array of] CDF non-varying field names.
; TAGNAMES: optional array of structure tag names.
; NVTAGNAMES: optional array of non-varying structure tag names.
; RESOLUTION: resolution to return in seconds.
; APPEND: if set, append data to the end of x.
; TIME: If set, will create tag TIME using the Epoch variable.
;SEE ALSO:
; "loadcdf2", "loadallcdf", "print_cdf_info","make_cdf_index"
;
;CREATED BY: Davin Larson
;MODIFICATIONS:
;LAST MODIFICATION: %W% %E%
;-
pro loadcdfstr,data0,novardata $
,filenames=cdf_files $
,path=path $
,varnames=cdf_vars,tagnames=tagnames $
,novarnames=novarnames $
,resolution = res $
,median = med $
,filter_proc = filter_proc $
,append=append,time=time, nvtagnames=nvtagnames $
,novarznames=novarznames, nvztagnames=nvztagnames
if n_elements(cdf_files) eq 0 then cdf_files=pickfile(filter="*.cdf",path=path,get_path=path)
;on_ioerror,skip
if keyword_set(append) then index=n_elements(data0)
for num = 0,n_elements(cdf_files)-1 do begin
cdf_file = cdf_files(num)
id = 0
if data_type(cdf_file) eq 7 then id = cdf_open(cdf_file) else id = cdf_file
if not keyword_set(silent) then print,'Loading ',cdf_file
inq = cdf_info(id)
if not keyword_set(cdf_vars) then begin ;get only variables that vary
for n=0,inq.nvars-1 do begin
vinq=cdf_varinq(id,n)
if vinq.recvar eq 'VARY' then append_array,cdf_vars,vinq.name
endfor
for n=0,inq.nzvars-1 do begin
vinq=cdf_varinq(id,n,/zvar)
if vinq.recvar eq 'VARY' then append_array,cdf_vars,vinq.name
endfor
endif
nvars = n_elements(cdf_vars)
if not keyword_set(tagnames) then tagnames = cdf_vars
tagnames = strcompress(tagnames,/remove_all)
if not keyword_set(data0) then append=0
if not keyword_set(append) then begin ;define the data structure:
; if keyword_set(time) then dat = {TIME:0.d}
for n=0,nvars-1 do begin
vinq = cdf_varinq(id,cdf_vars(n))
if vinq.is_zvar then dim = vinq.dim else dim = inq.dim*vinq.dimvar
w = where(dim,ndim)
if ndim gt 0 then dim=dim(w) else dim=0
;print,cdf_vars(n),ndim,dim
case vinq.datatype of
'CDF_REAL8' : value = !values.d_nan
'CDF_DOUBLE': value = !values.d_nan
'CDF_REAL4' : value = !values.f_nan
'CDF_FLOAT' : value = !values.f_nan
'CDF_INT4' : value = 0l
'CDF_UINT4' : value = 0ul
'CDF_INT2' : value = 0
'CDF_UINT2' : value = 0
'CDF_INT1' : value = 0b
'CDF_UINT1' : value = 0b
'CDF_CHAR' : value = 0b
'CDF_UCHAR' : value = 0b
'CDF_BYTE' : value = 0b
'CDF_EPOCH' : value = !values.d_nan
else : message ,'Invalid type, please fix source...'
endcase
if ndim gt 0 then val = make_array(value=value,dim=dim) $
else val=value
a = strpos(tagnames(n),'%')
aa = strpos(tagnames(n),'*')
if a ne -1 then begin
b = strlen(tagnames(n))
oldname = tagnames(n)
tagnames(n) = strmid(oldname,0,a)+'q'+strmid(oldname,a+1,b)
endif
if aa ne -1 then begin
b = strlen(tagnames(n))
oldname = tagnames(n)
tagnames(n) = strmid(oldname,0,aa)+'x'+strmid(oldname,aa+1,b)
endif
str_element,/add,dat,tagnames(n),val
endfor
if keyword_set(time) then begin
w = where(tag_names(dat) eq 'TIME',c)
if c eq 0 then str_element,/add,dat,'TIME',0.d
endif
endif else dat = data0(0)
vinq = cdf_varinq(id,cdf_vars(0))
!quiet = 1
cdf_control,id,variable=cdf_vars(0),get_var_info=varinfo,zvar=vinq.is_zvar
!quiet = 0
nrecs = varinfo.maxrec+1
data = replicate(dat,nrecs)
del = 0
if keyword_set(time) then begin
if cdf_attexists(id,'DEPEND_0',cdf_vars(0),zvar=vinq.is_zvar) then $
cdf_attget,id,'DEPEND_0',cdf_vars(0),epochnum,zvar=vinq.is_zvar $
else begin
for thisvar=0,inq.nvars-1 do begin
vinq = cdf_varinq(id,thisvar)
if vinq.datatype eq 'CDF_EPOCH' then $
epochnum=vinq.name
endfor
if n_elements(epochnum) eq 0 then begin
for thisvar=0,inq.nzvars-1 do begin
vinq = cdf_varinq(id,thisvar,/zvar)
if vinq.datatype eq 'CDF_EPOCH' then $
epochnum=vinq.name
endfor
endif
endelse
loadcdf2,id,epochnum,x
epoch0 = 719528.d * 24.* 3600. * 1000. ;Jan 1, 1970
data.time = (x - epoch0)/1000.
endif
for n=0,nvars-1 do begin
if cdf_attexists(id,'DEPEND_0',cdf_vars(n),zvar=vinq.is_zvar) then $
cdf_attget,id,'DEPEND_0',cdf_vars(n),thisepoch,zvar=vinq.is_zvar $
else thisepoch = epochnum
if n eq 0 and not keyword_set(time) then epochnum = thisepoch
if strpos(tagnames(n),'Epoch') ne -1 then thisepoch = tagnames(n)
if thisepoch eq epochnum then begin
loadcdf2,id,cdf_vars(n),x,/no_shift,nrecs=nrecs
if cdf_attexists(id,'FILLVAL',cdf_vars(n),zvar=vinq.is_zvar) then $
begin
case data_type(x) of
4: nan = !values.f_nan
5: nan = !values.d_nan
else: nan = 0
endcase
if nan ne 0 then begin
cdf_attget,id,'FILLVAL',cdf_vars(n),fv,zvar=vinq.is_zvar
fvindx = where(x eq fv,fvcnt)
if fvcnt gt 0 then x(fvindx) = nan
endif
endif
str_element,/add,data,tagnames(n),x
endif else begin
print,'Variable '+cdf_vars(n)+' has different Epoch'
endelse
endfor
if num eq 0 and keyword_set(novarnames) then begin
novardata = 0
novartags = strcompress(novarnames,/remove_all)
for i=0,n_elements(novarnames)-1 do begin
loadcdf2,id,novarnames(i),val
if keyword_set(nvtagnames) then nvtag = nvtagnames[i] else $
nvtag = novartags(i)
str_element,/add,novardata,nvtag,val
endfor
endif
if num eq 0 and keyword_set(novarznames) then begin
novarzdata = 0
novarztags = strcompress(novarznames,/remove_all)
for i=0,n_elements(novarznames)-1 do begin
loadcdf2,id,novarznames(i),val,/zvar
if keyword_set(nvztagnames) then nvztag = nvztagnames[i] else $
nvztag = novarztags(i)
str_element,/add,novardata,nvztag,val
endfor
endif
if data_type(cdf_file) eq 7 then cdf_close,id
if keyword_set(filter_proc) then call_procedure,filter_proc,data
if keyword_set(res) then data = average_str(data,res,/nan,median=med)
append_array,data0,data,index=index
append = 1
skip:
if id eq 0 then print,'Unable to open file: ',cdf_file
endfor
append_array,data0,index=index,/done
end