forrest@0
|
1 |
;----------------------------------------------------------------------
|
forrest@0
|
2 |
; change Month to yyyymm
|
forrest@0
|
3 |
; add Month to output
|
forrest@0
|
4 |
; add lat, lon to output
|
forrest@0
|
5 |
; add year to output
|
forrest@0
|
6 |
;
|
forrest@0
|
7 |
; This example reads an ASCII file that is formatted a specific way, and
|
forrest@0
|
8 |
; writes out the results to a netCDF file.
|
forrest@0
|
9 |
;
|
forrest@0
|
10 |
; The first line in the ASCII file must be a header, with each field
|
forrest@0
|
11 |
; separated by a single character delimiter (like a ","). The rest of
|
forrest@0
|
12 |
; the file must be such that each row contains all fields, each
|
forrest@0
|
13 |
; separated by the designated delimiter.
|
forrest@0
|
14 |
;
|
forrest@0
|
15 |
; The fields can be integer, float, double, character, or string.
|
forrest@0
|
16 |
; String fields cannot be written to a netCDF file. They have to
|
forrest@0
|
17 |
; be read in as character arrays and written out that way.
|
forrest@0
|
18 |
;----------------------------------------------------------------------
|
forrest@0
|
19 |
|
forrest@0
|
20 |
;----------------------------------------------------------------------
|
forrest@0
|
21 |
; This function returns the index locations of the given delimiter
|
forrest@0
|
22 |
; in a row or several rows of strings.
|
forrest@0
|
23 |
;----------------------------------------------------------------------
|
forrest@0
|
24 |
function delim_indices(strings,nfields,delimiter)
|
forrest@0
|
25 |
local cstrings, cdelim
|
forrest@0
|
26 |
begin
|
forrest@0
|
27 |
nrows = dimsizes(strings)
|
forrest@0
|
28 |
;
|
forrest@0
|
29 |
; Handle special case if we only have one string. Make sure it
|
forrest@0
|
30 |
; is put into a 2D array.
|
forrest@0
|
31 |
;
|
forrest@0
|
32 |
if(nrows.eq.1) then
|
forrest@0
|
33 |
cstrings = new((/1,strlen(strings)+1/),character)
|
forrest@0
|
34 |
end if
|
forrest@0
|
35 |
|
forrest@0
|
36 |
cstrings = stringtochar(strings) ; Convert to characters.
|
forrest@0
|
37 |
cdelim = stringtochar(delimiter) ; Convert delimiter to character.
|
forrest@0
|
38 |
;
|
forrest@0
|
39 |
; Som error checking here. Make sure delimiter is one character.
|
forrest@0
|
40 |
;
|
forrest@0
|
41 |
nc = dimsizes(cdelim)
|
forrest@0
|
42 |
rank = dimsizes(nc)
|
forrest@0
|
43 |
if(rank.ne.1.or.(rank.eq.1.and.nc.ne.2)) then
|
forrest@0
|
44 |
print("delim_indices: fatal: the delimiter you've selected")
|
forrest@0
|
45 |
print("must be a single character. Can't continue.")
|
forrest@0
|
46 |
exit
|
forrest@0
|
47 |
end if
|
forrest@0
|
48 |
|
forrest@0
|
49 |
;
|
forrest@0
|
50 |
; Create array to hold indices of delimiter locations, and then loop
|
forrest@0
|
51 |
; through each row and find all the delimiters. Make sure each row has
|
forrest@0
|
52 |
; the correct number of delimiters.
|
forrest@0
|
53 |
;
|
forrest@0
|
54 |
ndelims = nfields-1
|
forrest@0
|
55 |
cindices = new((/nrows,ndelims/),integer)
|
forrest@0
|
56 |
do i = 0, nrows-1
|
forrest@0
|
57 |
ii = ind(cstrings(i,:).eq.cdelim(0))
|
forrest@0
|
58 |
;
|
forrest@0
|
59 |
; Make sure there were delimiters on this row. If not, we just quit.
|
forrest@0
|
60 |
; This could probably be modified to do this more gracefully.
|
forrest@0
|
61 |
;
|
forrest@0
|
62 |
if(any(ismissing(ii))) then
|
forrest@0
|
63 |
print("delim_indices: fatal: I didn't find any delimiters")
|
forrest@0
|
64 |
print("('" + delimiter + "') on row " + i + ". Can't continue.")
|
forrest@0
|
65 |
exit
|
forrest@0
|
66 |
end if
|
forrest@0
|
67 |
if(dimsizes(ii).ne.ndelims) then
|
forrest@0
|
68 |
print("delim_indices: fatal: I expected to find " + ndelims)
|
forrest@0
|
69 |
print("delimiters on row " + i + ". Instead, I found " + dimsizes(ii) + ".")
|
forrest@0
|
70 |
print("Can't continue.")
|
forrest@0
|
71 |
exit
|
forrest@0
|
72 |
end if
|
forrest@0
|
73 |
|
forrest@0
|
74 |
cindices(i,:) = ii
|
forrest@0
|
75 |
|
forrest@0
|
76 |
delete(ii) ; For next time through loop
|
forrest@0
|
77 |
end do
|
forrest@0
|
78 |
|
forrest@0
|
79 |
return(cindices)
|
forrest@0
|
80 |
end
|
forrest@0
|
81 |
|
forrest@0
|
82 |
;----------------------------------------------------------------------
|
forrest@0
|
83 |
; This function reads in a particular field from a string array,
|
forrest@0
|
84 |
; given the field number to read (fields start at #1 and go to #nfield),
|
forrest@0
|
85 |
; and the indices of the delimiters.
|
forrest@0
|
86 |
;
|
forrest@0
|
87 |
; It returns either an integer, float, double, character, or a string,
|
forrest@0
|
88 |
; depending on the input flag "return_type".
|
forrest@0
|
89 |
;----------------------------------------------------------------------
|
forrest@0
|
90 |
function read_field(strings,ifield,indices,return_type)
|
forrest@0
|
91 |
local nstring, cstrings, nf, tmp_str
|
forrest@0
|
92 |
begin
|
forrest@0
|
93 |
nrows = dimsizes(strings)
|
forrest@0
|
94 |
;
|
forrest@0
|
95 |
; Handle special case if we only have one string. Make sure it
|
forrest@0
|
96 |
; is put into a 2D array.
|
forrest@0
|
97 |
;
|
forrest@0
|
98 |
if(nrows.eq.1) then
|
forrest@0
|
99 |
cstrings = new((/1,strlen(strings)+1/),character)
|
forrest@0
|
100 |
end if
|
forrest@0
|
101 |
|
forrest@0
|
102 |
cstrings = stringtochar(strings)
|
forrest@0
|
103 |
nf = dimsizes(indices(0,:))+1 ; indices is nrows x (nfields-1)
|
forrest@0
|
104 |
|
forrest@0
|
105 |
;
|
forrest@0
|
106 |
; Error checking. Make sure user has entered a valid field.
|
forrest@0
|
107 |
;
|
forrest@0
|
108 |
if(ifield.le.0.or.ifield.gt.nf) then
|
forrest@0
|
109 |
print("read_field: fatal: you've selected a field that is")
|
forrest@0
|
110 |
print("out-of-range of the number of fields that you have (" + nf + ").")
|
forrest@0
|
111 |
exit
|
forrest@0
|
112 |
end if
|
forrest@0
|
113 |
|
forrest@0
|
114 |
;
|
forrest@0
|
115 |
; Set up array to return. For string, int, float, or double arrays,
|
forrest@0
|
116 |
; we don't have to do anything special. For character arrays,
|
forrest@0
|
117 |
; however, we do.
|
forrest@0
|
118 |
;
|
forrest@0
|
119 |
if(return_type.ne."character") then
|
forrest@0
|
120 |
return_array = new(nrows,return_type)
|
forrest@0
|
121 |
else
|
forrest@0
|
122 |
;
|
forrest@0
|
123 |
; We don't know what the biggest character array is at this point, so
|
forrest@0
|
124 |
; make it bigger than necessary, and then resize later as necessary.
|
forrest@0
|
125 |
;
|
forrest@0
|
126 |
tmp_return_array = new((/nrows,dimsizes(cstrings(0,:))/),"character")
|
forrest@0
|
127 |
|
forrest@0
|
128 |
max_len = 0 ; Use to keep track of max lengths of strings.
|
forrest@0
|
129 |
end if
|
forrest@0
|
130 |
|
forrest@0
|
131 |
do i = 0,nrows-1
|
forrest@0
|
132 |
;
|
forrest@0
|
133 |
; Special case of first field in row.
|
forrest@0
|
134 |
;
|
forrest@0
|
135 |
if(ifield.eq.1) then
|
forrest@0
|
136 |
ibeg = 0
|
forrest@0
|
137 |
iend = indices(i,ifield-1)-1
|
forrest@0
|
138 |
else
|
forrest@0
|
139 |
;
|
forrest@0
|
140 |
; Special case of first field in row.
|
forrest@0
|
141 |
;
|
forrest@0
|
142 |
if(ifield.eq.nf) then
|
forrest@0
|
143 |
ibeg = indices(i,ifield-2)+1
|
forrest@0
|
144 |
iend = dimsizes(cstrings(i,:))-1
|
forrest@0
|
145 |
;
|
forrest@0
|
146 |
; Any field between first and last field.
|
forrest@0
|
147 |
;
|
forrest@0
|
148 |
else
|
forrest@0
|
149 |
ibeg = indices(i,ifield-2)+1
|
forrest@0
|
150 |
iend = indices(i,ifield-1)-1
|
forrest@0
|
151 |
end if
|
forrest@0
|
152 |
end if
|
forrest@0
|
153 |
;
|
forrest@0
|
154 |
; Here's the code that pulls off the correct string, and converts it
|
forrest@0
|
155 |
; to float if desired.
|
forrest@0
|
156 |
;
|
forrest@0
|
157 |
if(return_type.eq."integer") then
|
forrest@0
|
158 |
return_array(i) = stringtointeger(chartostring(cstrings(i,ibeg:iend)))
|
forrest@0
|
159 |
end if
|
forrest@0
|
160 |
if(return_type.eq."float") then
|
forrest@0
|
161 |
return_array(i) = stringtofloat(chartostring(cstrings(i,ibeg:iend)))
|
forrest@0
|
162 |
end if
|
forrest@0
|
163 |
if(return_type.eq."double") then
|
forrest@0
|
164 |
return_array(i) = stringtodouble(chartostring(cstrings(i,ibeg:iend)))
|
forrest@0
|
165 |
end if
|
forrest@0
|
166 |
if(return_type.eq."string") then
|
forrest@0
|
167 |
return_array(i) = chartostring(cstrings(i,ibeg:iend))
|
forrest@0
|
168 |
end if
|
forrest@0
|
169 |
if(return_type.eq."character") then
|
forrest@0
|
170 |
if( (iend-ibeg+1) .gt. max_len) then
|
forrest@0
|
171 |
max_len = iend-ibeg+1
|
forrest@0
|
172 |
end if
|
forrest@0
|
173 |
tmp_return_array(i,0:iend-ibeg) = cstrings(i,ibeg:iend)
|
forrest@0
|
174 |
end if
|
forrest@0
|
175 |
end do
|
forrest@0
|
176 |
|
forrest@0
|
177 |
if(return_type.eq."character") then
|
forrest@0
|
178 |
return_array = new((/nrows,max_len/),"character")
|
forrest@0
|
179 |
return_array = tmp_return_array(:,0:max_len-1)
|
forrest@0
|
180 |
end if
|
forrest@0
|
181 |
|
forrest@0
|
182 |
return(return_array)
|
forrest@0
|
183 |
end
|
forrest@0
|
184 |
|
forrest@0
|
185 |
|
forrest@0
|
186 |
;----------------------------------------------------------------------
|
forrest@0
|
187 |
; This function reads in string fields only to get the maximum string
|
forrest@0
|
188 |
; length.
|
forrest@0
|
189 |
;----------------------------------------------------------------------
|
forrest@0
|
190 |
function get_maxlen(strings,ifield,indices)
|
forrest@0
|
191 |
local nstring, cstrings, nf, tmp_str
|
forrest@0
|
192 |
begin
|
forrest@0
|
193 |
nrows = dimsizes(strings)
|
forrest@0
|
194 |
;
|
forrest@0
|
195 |
; Handle special case if we only have one string. Make sure it
|
forrest@0
|
196 |
; is put into a 2D array.
|
forrest@0
|
197 |
;
|
forrest@0
|
198 |
if(nrows.eq.1) then
|
forrest@0
|
199 |
cstrings = new((/1,strlen(strings)+1/),character)
|
forrest@0
|
200 |
end if
|
forrest@0
|
201 |
|
forrest@0
|
202 |
cstrings = stringtochar(strings)
|
forrest@0
|
203 |
nf = dimsizes(indices(0,:))+1 ; indices is nrows x (nfields-1)
|
forrest@0
|
204 |
|
forrest@0
|
205 |
;
|
forrest@0
|
206 |
; Error checking. Make sure user has entered a valid field.
|
forrest@0
|
207 |
;
|
forrest@0
|
208 |
if(ifield.le.0.or.ifield.gt.nf) then
|
forrest@0
|
209 |
print("read_field: fatal: you've selected a field that is")
|
forrest@0
|
210 |
print("out-of-range of the number of fields that you have (" + nf + ").")
|
forrest@0
|
211 |
exit
|
forrest@0
|
212 |
end if
|
forrest@0
|
213 |
;
|
forrest@0
|
214 |
; We don't know what the biggest character array is at this point, so
|
forrest@0
|
215 |
; make it bigger than necessary, and then resize later as necessary.
|
forrest@0
|
216 |
;
|
forrest@0
|
217 |
tmp_return_array = new((/nrows,dimsizes(cstrings(0,:))/),"character")
|
forrest@0
|
218 |
|
forrest@0
|
219 |
max_len = 0 ; Use to keep track of max lengths of strings.
|
forrest@0
|
220 |
|
forrest@0
|
221 |
do i = 0,nrows-1
|
forrest@0
|
222 |
;
|
forrest@0
|
223 |
; Special case of first field in row.
|
forrest@0
|
224 |
;
|
forrest@0
|
225 |
if(ifield.eq.1) then
|
forrest@0
|
226 |
ibeg = 0
|
forrest@0
|
227 |
iend = indices(i,ifield-1)-1
|
forrest@0
|
228 |
else
|
forrest@0
|
229 |
;
|
forrest@0
|
230 |
; Special case of first field in row.
|
forrest@0
|
231 |
;
|
forrest@0
|
232 |
if(ifield.eq.nf) then
|
forrest@0
|
233 |
ibeg = indices(i,ifield-2)+1
|
forrest@0
|
234 |
iend = dimsizes(cstrings(i,:))-1
|
forrest@0
|
235 |
;
|
forrest@0
|
236 |
; Any field between first and last field.
|
forrest@0
|
237 |
;
|
forrest@0
|
238 |
else
|
forrest@0
|
239 |
ibeg = indices(i,ifield-2)+1
|
forrest@0
|
240 |
iend = indices(i,ifield-1)-1
|
forrest@0
|
241 |
end if
|
forrest@0
|
242 |
end if
|
forrest@0
|
243 |
if( (iend-ibeg+1) .gt. max_len) then
|
forrest@0
|
244 |
max_len = iend-ibeg+1
|
forrest@0
|
245 |
end if
|
forrest@0
|
246 |
end do
|
forrest@0
|
247 |
|
forrest@0
|
248 |
return(max_len)
|
forrest@0
|
249 |
end
|
forrest@0
|
250 |
|
forrest@0
|
251 |
;----------------------------------------------------------------------
|
forrest@0
|
252 |
; Main code.
|
forrest@0
|
253 |
;----------------------------------------------------------------------
|
forrest@0
|
254 |
begin
|
forrest@0
|
255 |
|
forrest@0
|
256 |
;###############################################################
|
forrest@0
|
257 |
; Set up defaults. We are hard-coding here..
|
forrest@0
|
258 |
|
forrest@0
|
259 |
year = 1991
|
forrest@0
|
260 |
station = "USHa1"
|
forrest@0
|
261 |
lat = 42.5378
|
forrest@0
|
262 |
lon = -72.1715 + 360.
|
forrest@0
|
263 |
|
forrest@0
|
264 |
nfields = 30 ; # of fields
|
forrest@0
|
265 |
delimiter = "," ; field delimiter
|
forrest@0
|
266 |
|
forrest@0
|
267 |
filename = station+year+"_L4_m.txt" ; ASCII" file to read.
|
forrest@0
|
268 |
cdf_file = station+year+"_L4_m.nc" ; netCDF file to write.
|
forrest@0
|
269 |
|
forrest@0
|
270 |
if(isfilepresent(cdf_file))
|
forrest@0
|
271 |
print("Warning: '" + cdf_file + "' exists. Will remove it.")
|
forrest@0
|
272 |
system("/bin/rm " + cdf_file)
|
forrest@0
|
273 |
end if
|
forrest@0
|
274 |
|
forrest@0
|
275 |
; In this case, fields #1-#2 are integers,
|
forrest@0
|
276 |
; and the rest of the fields are floats.
|
forrest@0
|
277 |
|
forrest@0
|
278 |
var_types = new(nfields,string)
|
forrest@0
|
279 |
var_types = "float" ; Most are floats.
|
forrest@0
|
280 |
var_types(0:1) = "integer"
|
forrest@0
|
281 |
|
forrest@0
|
282 |
;#####################################################################
|
forrest@0
|
283 |
|
forrest@0
|
284 |
; Read in data as strings. This will create a string array that has the
|
forrest@0
|
285 |
; same number of strings as there are rows in the file. We will then need
|
forrest@0
|
286 |
; to parse each string later.
|
forrest@0
|
287 |
|
forrest@0
|
288 |
read_data = asciiread(filename,-1,"string")
|
forrest@0
|
289 |
|
forrest@0
|
290 |
header = read_data(0) ; Header. Use for variable names.
|
forrest@0
|
291 |
data = read_data(1:) ; Get rid of first line which is a header.
|
forrest@0
|
292 |
nmonth = dimsizes(data) ; Number of rows == number of month.
|
forrest@0
|
293 |
|
forrest@0
|
294 |
; Read in locations of delimiters in each string row.
|
forrest@0
|
295 |
|
forrest@0
|
296 |
hindices = delim_indices(header,nfields,delimiter) ; header row
|
forrest@0
|
297 |
dindices = delim_indices(data,nfields,delimiter) ; rest of file
|
forrest@0
|
298 |
|
forrest@0
|
299 |
; Read in the field names which will become variable names on
|
forrest@0
|
300 |
; the netCDF file.
|
forrest@0
|
301 |
|
forrest@0
|
302 |
var_names = new(nfields,string)
|
forrest@0
|
303 |
|
forrest@0
|
304 |
do i=0,nfields-1
|
forrest@0
|
305 |
var_names(i) = read_field(header,i+1,hindices,"string")
|
forrest@0
|
306 |
end do
|
forrest@0
|
307 |
|
forrest@0
|
308 |
;-------------------------------------------------------------------
|
forrest@0
|
309 |
; Write out this netCDF file efficiently so it will be faster.
|
forrest@0
|
310 |
; Try to predefine everything before you write to it.
|
forrest@0
|
311 |
|
forrest@0
|
312 |
f = addfile(cdf_file,"c")
|
forrest@0
|
313 |
setfileoption(f,"DefineMode",True) ; Enter predefine phase.
|
forrest@0
|
314 |
|
forrest@0
|
315 |
; Write global attributes to file. It's okay to do this before
|
forrest@0
|
316 |
; predefining the file's variables. We are still in "define" mode.
|
forrest@0
|
317 |
|
forrest@0
|
318 |
fAtt = True
|
forrest@0
|
319 |
fAtt@description = "Data read in from " + filename + " ASCII file."
|
forrest@0
|
320 |
fAtt@creation_date = systemfunc ("date")
|
forrest@0
|
321 |
fileattdef( f, fAtt )
|
forrest@0
|
322 |
|
forrest@0
|
323 |
; Write dimension names to file.
|
forrest@0
|
324 |
|
forrest@0
|
325 |
dim_names = (/ "year", "month" /)
|
forrest@0
|
326 |
dim_sizes = (/ -1 , nmonth /)
|
forrest@0
|
327 |
dimUnlim = (/ True , False /)
|
forrest@0
|
328 |
filedimdef( f, dim_names, dim_sizes, dimUnlim )
|
forrest@0
|
329 |
|
forrest@0
|
330 |
filedimdef( f, "lat", 1, False )
|
forrest@0
|
331 |
filedimdef( f, "lon", 1, False )
|
forrest@0
|
332 |
|
forrest@0
|
333 |
; Define each variable on the file.
|
forrest@0
|
334 |
|
forrest@0
|
335 |
filevardef( f, "year", "integer", "year" )
|
forrest@0
|
336 |
filevardef( f, "lat" , "float" , "lat" )
|
forrest@0
|
337 |
filevardef( f, "lon" , "float" , "lon" )
|
forrest@0
|
338 |
|
forrest@0
|
339 |
do i=0,nfields-1
|
forrest@0
|
340 |
filevardef(f, var_names(i), var_types(i), dim_names)
|
forrest@0
|
341 |
end do
|
forrest@0
|
342 |
;-----------------------------------------------------------------
|
forrest@0
|
343 |
|
forrest@0
|
344 |
; Loop through each field, read the values for that field, print
|
forrest@0
|
345 |
; information about the variable, and then write it to the netCDF file.
|
forrest@0
|
346 |
|
forrest@0
|
347 |
do i=0,nfields-1
|
forrest@0
|
348 |
ifield = i+1 ; Fields start at #1, not #0.
|
forrest@0
|
349 |
|
forrest@0
|
350 |
tmp_data = new((/1,nmonth/),var_types(i))
|
forrest@0
|
351 |
|
forrest@0
|
352 |
if (i.le.1) then
|
forrest@0
|
353 |
tmp_data@_FillValue = -999
|
forrest@0
|
354 |
else
|
forrest@0
|
355 |
tmp_data@_FillValue = -9999.00
|
forrest@0
|
356 |
end if
|
forrest@0
|
357 |
|
forrest@0
|
358 |
tmp_data(0,:) = read_field(data,ifield,dindices,var_types(i))
|
forrest@0
|
359 |
|
forrest@0
|
360 |
; change Month to yyyymm
|
forrest@0
|
361 |
|
forrest@0
|
362 |
if (i.eq.0) then
|
forrest@0
|
363 |
tmp_data(0,:) = tmp_data(0,:) + year*100
|
forrest@0
|
364 |
end if
|
forrest@0
|
365 |
|
forrest@0
|
366 |
; Print some info about the variable.
|
forrest@0
|
367 |
|
forrest@0
|
368 |
; print("")
|
forrest@0
|
369 |
; print("Writing variable '" + var_names(i) + "' (field #" + ifield + ").")
|
forrest@0
|
370 |
; print("Type is " + var_types(i) + ".")
|
forrest@0
|
371 |
; print("min/max = " + min(tmp_data) + "/" + max(tmp_data))
|
forrest@0
|
372 |
|
forrest@0
|
373 |
; if(any(ismissing(tmp_data))) then
|
forrest@0
|
374 |
; print("This variable does contain missing values.")
|
forrest@0
|
375 |
; else
|
forrest@0
|
376 |
; print("This variable doesn't contain missing values.")
|
forrest@0
|
377 |
; end if
|
forrest@0
|
378 |
|
forrest@0
|
379 |
; write variable to file
|
forrest@0
|
380 |
|
forrest@0
|
381 |
f->$var_names(i)$ = tmp_data ; Write to netCDF file.
|
forrest@0
|
382 |
|
forrest@0
|
383 |
delete(tmp_data) ; Delete for next round.
|
forrest@0
|
384 |
end do
|
forrest@0
|
385 |
|
forrest@0
|
386 |
; write variable to file
|
forrest@0
|
387 |
|
forrest@0
|
388 |
f->year = year
|
forrest@0
|
389 |
f->lat = lat
|
forrest@0
|
390 |
f->lon = lon
|
forrest@0
|
391 |
end
|