bug fixed in whitelist sorting, more robust against numpy parsing errors

This commit is contained in:
Philip Eisenlohr 2017-08-28 19:38:14 -04:00
parent 5b8bc65d37
commit 64259d9239
1 changed files with 12 additions and 7 deletions

View File

@ -93,8 +93,8 @@ for name in filenames:
or fnmatch.fnmatch(label,needle) for needle in options.whitelist] # which whitelist items do match it
whitelistitem[i] = match.index(True) if np.sum(match) == 1 else -1 # unique match to a whitelist item --> store which
sorted = np.lexsort(sortingList(labels,whitelistitem))
order = range(len(labels)) if sorted[0] < 0 else sorted # skip reordering if non-unique, i.e. first sorted is "-1"
order = range(len(labels)) if np.any(whitelistitem < 0) \
else np.lexsort(sortingList(labels,whitelistitem)) # reorder if unique, i.e. no "-1" in whitelistitem
else:
order = range(len(labels)) # maintain original order of labels
@ -133,14 +133,19 @@ for name in filenames:
positions = np.array(positions)[order]
if options.condition is None: # read full array and filter columns
table.data_readArray(1+positions) # read desired columns (indexed 1,...)
table.data_writeArray() # directly write out
else:
atOnce = options.condition is None
if atOnce: # read full array and filter columns
try:
table.data_readArray(positions+1) # read desired columns (indexed 1,...)
table.data_writeArray() # directly write out
except:
atOnce = False # data contains items that prevent array chunking
if not atOnce: # read data line by line
outputAlive = True
while outputAlive and table.data_read(): # read next data line of ASCII table
specials['_row_'] += 1 # count row
if eval(condition): # valid row ?
if options.condition is None or eval(condition): # valid row ?
table.data = [table.data[position] for position in positions] # retain filtered columns
outputAlive = table.data_write() # output processed line