issues simply parsing a whitespace-delimited textfile in pythonscript

[Damon Getsman]

Okay so I'm writing a script in python right now as a dirty fix for a
problem we're having at work.. Unfortunately this is the first really
non-trivial script that I've had to work with in python and the book
that I have on it really kind of sucks.

I'm having an issue parsing lines of 'last' output that I have stored
in a /tmp file. The first time it does a .readline() I get the full
line of output, which I'm then able to split() and work with the
individual fields of without any problem. Unfortunately, the second
time that I do a .readline() on the file, I am only receiving the
first character of the first field. Looking through the /tmp file
shows that it's not corrupted from the format that it should be in at
all... Here's the relevant script:

----
#parse
Lastdump = open('/tmp/esd_tmp', 'r')

#find out what the last day entry is in the wtmp
cur_rec = Lastdump.readline()
work = cur_rec.split()

if debug == 1:
print work
print " is our split record line from /tmp/esd_tmp\n"

startday = work[3]

if debug == 1:
print startday + " is the starting day\n"
print days
print " is our dictionary of days\n"
print days[startday] + " is our ending day\n"

for cur_rec in Lastdump.readline():
work = cur_rec.split()

if debug == 1:
print "Starting table building pass . . .\n"
print work
print " is the contents of our split record line now\n"
print cur_rec + " is the contents of cur_rec\n"

#only go back 2 days

while work[0] != days[startday]:
tmp = work[1]
if table.has_key(work[0]):
continue
elif tmp[0] != ':':
#don't keep it if it isn't a SunRay terminal
identifier
continue
else:
#now we keep it
table[work[0]] = tmp
----

the first and second sets of debugging output show everything as they
should be... the third shows that the next working line (in cur_rec),
and thus 'work', as well, only hold the first character of the line.
Here's the output:

----
Debugging run
Building table . . .

['dgetsman', 'pts/3', ':0.0', 'Wed', 'May', '21', '10:21', 'still',
'logged',
'in']
is our split record line from /tmp/esd_tmp

Wed is the starting day

{'Wed': 'Mon', 'Sun': 'Fri', 'Fri': 'Wed', 'Thurs': 'Tues', 'Tues':
'Sun',
'Mon': 'Sat', 'Sat': 'Thurs'}
is our dictionary of days

Mon is our ending day

Starting table building pass . . .

['d']
is the contents of our split record line now

d is the contents of cur_rec

----
And thus everything fails when I try to work with the different fields
in subsequent script afterwards. Does anybody have an idea as to why
this would be happening?

Oh, and if relevant, here's the datafile's first few lines:

----
dgetsman pts/3 :0.0 Wed May 21 10:21 still logged
in
dgetsman pts/2 :0.0 Wed May 21 09:04 still logged
in
dgetsman pts/1 :0.0 Wed May 21 08:56 - 10:21
(01:24)
dgetsman pts/0 :0.0 Wed May 21 08:56 still logged
in

I would really appreciate any pointers or suggestions you can give.

<a href="http://www.zoominfo.com/people/Getsman_Damon_-214241.aspx">
*Damon Getsman
Linux/Solaris System Administrator
</a>

[Damon Getsman]

Okay, so I manged to kludge around the issue by not using
the .readline() in my 'for' statement. Instead, I'm slurping the
whole file into a new list that I put in for that purpose, and
everything seems to be working just fine. However, I don't know WHY
the other method failed and I'm at a loss for why that didn't work and
this is working. I'd really like to know the why about this issue so
that I don't have to use crappy coding practice and kludge around it
the next time I have an issue like this.

Any ideas much appreciated.

Damon G.

[Paul McGuire]

On May 21, 10:59*am, Damon Getsman <[e-mail]:

I'm having an issue parsing lines of 'last' output that I have stored
in a /tmp file. *The first time it does a .readline() I get the full
line of output, which I'm then able to split() and work with the
individual fields of without any problem. *Unfortunately, the second
time that I do a .readline() on the file, I am only receiving the
first character of the first field. *Looking through the /tmp file
shows that it's not corrupted from the format that it should be in at
all... *Here's the relevant script:

----
* * #parse
* * Lastdump = open('/tmp/esd_tmp', 'r')

* * #find out what the last day entry is in the wtmp
* * cur_rec = Lastdump.readline()
* * work = cur_rec.split()

* * if debug == 1:
* * * * print work
* * * * print " is our split record line from /tmp/esd_tmp\n"

* * startday = work[3]

* * if debug == 1:
* * * * print startday + " is the starting day\n"
* * * * print days
* * * * print " is our dictionary of days\n"
* * * * print days[startday] + " is our ending day\n"

* * for cur_rec in Lastdump.readline():
* * * * work = cur_rec.split()

<snip>
for cur_rec in Lastdump.readline():

is the problem. readline() returns a string containing the next
line's worth of text, NOT an iterator over all the subsequent lines in
the file. So your code is really saying:

next_line_in_file = Lastdump.readline():
for cur_rec in next_line_in_file:

which of course, is iterating over a string character by character.

Since you are opening Lastdump (not great casing for a variable name,
BTW - looks like a class name with that leading capital letter), it
gives you an iterator already. Try this instead:

lastdump = open('/tmp/esd_tmp', 'r')

cur_rec = lastdump.next()

...

for cur_rec in lastdump:

...

This should get you over the hump on reading the file.

Also, may I suggest this method for splitting up each record line, and
assigning individual fields to variables:

user,s1,s2,day,month,date,time,desc = cur_rec.split(None,7)

-- Paul

[Damon Getsman]

On May 21, 11:15 am, Paul McGuire <[e-mail]:

<snip>

for cur_rec in Lastdump.readline():

is the problem. readline() returns a string containing the next
line's worth of text, NOT an iterator over all the subsequent lines in
the file. So your code is really saying:

next_line_in_file = Lastdump.readline():
for cur_rec in next_line_in_file:

which of course, is iterating over a string character by character.

Since you are opening Lastdump (not great casing for a variable name,
BTW - looks like a class name with that leading capital letter), it
gives you an iterator already. Try this instead:

lastdump = open('/tmp/esd_tmp', 'r')

cur_rec = lastdump.next()

...

for cur_rec in lastdump:

...

This should get you over the hump on reading the file.

Also, may I suggest this method for splitting up each record line, and
assigning individual fields to variables:

user,s1,s2,day,month,date,time,desc = cur_rec.split(None,7)

-- Paul

Well the individual variables isn't exactly appropriate as I'm only
going to be using 2 of the fields. I think I will set those to
individual variables with a slice of what you mentioned, though, for
readability. Thank you for the tips, they were all much appreciated.

-Damon