PyFileSlice
From Sfvlug
(Difference between revisions)
(Source code to a fileslicing program in python) |
|||
| Line 48: | Line 48: | ||
for line in awsdata: | for line in awsdata: | ||
print line | print line | ||
| + | |||
| + | |||
| + | Brian and I had some disagreement about the use of the ''pass'' keyword in Python. I'm of the mind that pass is a similar analog to goto/break/etc. that should be avoided. Research has shown that use of ''pass'' is a normal Python idiom however. In this case, it actually makes the code cleaner, since it's simply slicing through a properly delimited text file. | ||
| + | |||
| + | However, below is an alternate version that uses nested loop and a boolean to signal when to exit the loop. In a C++ world, this is technically the "right" way to do it, but in Python, I'm not so sure. The ''pass'' version is much easier to understand. | ||
| + | |||
| + | signal = True | ||
| + | while signal == True: | ||
| + | line = infile.readline() | ||
| + | if line.startswith( "BEGIN_PAGEREFS"): | ||
| + | line = infile.readline() | ||
| + | while not line.startswith("END_PAGEREFS"): | ||
| + | tempLine = line.split(' ') | ||
| + | # note that the use of search() and more complete patterns | ||
| + | # might be more efficient despite compiling. | ||
| + | if re.search( pat1, tempLine[0] ): | ||
| + | awsParsed.append( [ tempLine[0], tempLine[1] ] ) | ||
| + | elif re.search( pat2, tempLine[0] ): | ||
| + | awsParsed.append( [ tempLine[0], tempLine[1] ] ) | ||
| + | line=infile.readline()[:-1] # remove trailing \ | ||
| + | , similar to chomp in perl. | ||
| + | signal = False | ||
| + | else: | ||
| + | line = infile.readline() | ||
Revision as of 01:52, 12 December 2006
#!/usr/bin/env python
#
# Simple tool to spit out referrer information from an awstats database
# for later searching an analysis. A good example of file slicing!
__author__ = "Nick Guy & Brian Guy"
__license__ = "GPL"
import sys, string;
# lolz, no argc it seems. :P
argc = len(sys.argv)
if argc > 2 :
print sys.argv[0] + " [filename]"
print "[filename] is optional, leave out to use stdin"
sys.exit(1)
# variables instantiated here to keep them in file scope.
awsdata = []
infile = False
if argc == 2:
try:
infile = open( sys.argv[1], 'r' )
except IOError:
print "Can't open " + sys.argv[1] + " for reading."
sys.exit(2)
if argc == 1:
infile = sys.stdin
# fastest method. Note that the strings inside startswith() are
# the start and end block tokens we need. Note also that the strings
# used to delimit the block we want are NOT included in the final output.
while not infile.readline().startswith("BEGIN_PAGEREFS"):
pass
# This is a syntactic hack to implement do/while loops.
line=infile.readline()
while not line.startswith("END_PAGEREFS"):
awsdata.append(line)
line=infile.readline()[:-1] # remove trailing \
, similar to chomp in perl.
infile.close() # send data to stdout. for line in awsdata: print line
Brian and I had some disagreement about the use of the pass keyword in Python. I'm of the mind that pass is a similar analog to goto/break/etc. that should be avoided. Research has shown that use of pass is a normal Python idiom however. In this case, it actually makes the code cleaner, since it's simply slicing through a properly delimited text file.
However, below is an alternate version that uses nested loop and a boolean to signal when to exit the loop. In a C++ world, this is technically the "right" way to do it, but in Python, I'm not so sure. The pass version is much easier to understand.
signal = True
while signal == True:
line = infile.readline()
if line.startswith( "BEGIN_PAGEREFS"):
line = infile.readline()
while not line.startswith("END_PAGEREFS"):
tempLine = line.split(' ')
# note that the use of search() and more complete patterns
# might be more efficient despite compiling.
if re.search( pat1, tempLine[0] ):
awsParsed.append( [ tempLine[0], tempLine[1] ] )
elif re.search( pat2, tempLine[0] ):
awsParsed.append( [ tempLine[0], tempLine[1] ] )
line=infile.readline()[:-1] # remove trailing \
, similar to chomp in perl.
signal = False else: line = infile.readline()
