Randyb
asked on
Filtering leading spaces from text string.
I need a quick fix and am experiencing brain fade so I thought I would offer you nice people a chance at some easy point, It's been a while since I posted a question so I thought it would be fun.
OK here ya go ... shouldnt be too tough.
Nightly we dump logs of our database and filesystem backups to a text "log" file. for one field entry we tail the "fixed" output file from Informix and pipe it through "head -1" to grab just the first line. So far this has worked fine but now I am formatting (mostly cleaning up) the format to be place in a database for query level reporting.
So...we have one line of text in one file, unfortunately this text has LEADING spaces that I would like to eliminate, theres your question.
How do I filter a line that looks something like " 143121 - 143856" into something like "143121 - 143856" ?
A couple things... first notice that there MAY be additional spaces embedded in the string before and after the "-" these can be left in place or filtered out, I dont care, what I do care about is that the actual text value (Database logical log numbers) stay in tact an accurate. ALSO the solution should be shell independent as we may use C shell on one system and K shell or Bourne shell on others. Thinking maybe AWK or SED or even a GREP statement might do it ???
Current syntax (if it helps) is:
(filename is an example)
LOGS="'tail -3 /testfile.flg | head -1 '"
if [ -f "/testfile.log" ]
then
ORIGNUM= 'head -1 "/testfile.log'
ORIGNUM1='expr $ORIGNUM -1'
ORIG=1
then this string gets echoed out to sendmail and to a logfile.
Its a little rough, since you dont have all the varible info but I think you see the drift.
I put the value at 100 cuz I feel like giving away points , sound fair ?
Anyone ?
THANX !!
OK here ya go ... shouldnt be too tough.
Nightly we dump logs of our database and filesystem backups to a text "log" file. for one field entry we tail the "fixed" output file from Informix and pipe it through "head -1" to grab just the first line. So far this has worked fine but now I am formatting (mostly cleaning up) the format to be place in a database for query level reporting.
So...we have one line of text in one file, unfortunately this text has LEADING spaces that I would like to eliminate, theres your question.
How do I filter a line that looks something like " 143121 - 143856" into something like "143121 - 143856" ?
A couple things... first notice that there MAY be additional spaces embedded in the string before and after the "-" these can be left in place or filtered out, I dont care, what I do care about is that the actual text value (Database logical log numbers) stay in tact an accurate. ALSO the solution should be shell independent as we may use C shell on one system and K shell or Bourne shell on others. Thinking maybe AWK or SED or even a GREP statement might do it ???
Current syntax (if it helps) is:
(filename is an example)
LOGS="'tail -3 /testfile.flg | head -1 '"
if [ -f "/testfile.log" ]
then
ORIGNUM= 'head -1 "/testfile.log'
ORIGNUM1='expr $ORIGNUM -1'
ORIG=1
then this string gets echoed out to sendmail and to a logfile.
Its a little rough, since you dont have all the varible info but I think you see the drift.
I put the value at 100 cuz I feel like giving away points , sound fair ?
Anyone ?
THANX !!
For 100 points, maybe I should have said "specification", not "spcification"
.
.
ORIGNUM=`head -1 /testfile.log|sed s/ //g`
.
.
Should there really have been a " before the /testfile.log in this line or was that a typo?
.
.
ORIGNUM=`head -1 /testfile.log|sed s/ //g`
.
.
Should there really have been a " before the /testfile.log in this line or was that a typo?
ASKER
Let me test drive this theory of your real quick
Actually I just figured it out also using sed but my solution has more arguments...I'll post it for you in a sec . If yours works and since its more "condensed" i'll give you the points, if not I will retract the question as my solution does work already.
be right back !
Actually I just figured it out also using sed but my solution has more arguments...I'll post it for you in a sec . If yours works and since its more "condensed" i'll give you the points, if not I will retract the question as my solution does work already.
be right back !
ASKER
Let me test drive this theory of your real quick
Actually I just figured it out also using sed but my solution has more arguments...I'll post it for you in a sec . If yours works and since its more "condensed" i'll give you the points, if not I will retract the question as my solution does work already.
be right back !
Actually I just figured it out also using sed but my solution has more arguments...I'll post it for you in a sec . If yours works and since its more "condensed" i'll give you the points, if not I will retract the question as my solution does work already.
be right back !
Seems fair, if you'll show me yours...
However, I reckon I thought of the answer before you did (see the timestamps on this Q...!)
However, I reckon I thought of the answer before you did (see the timestamps on this Q...!)
ASKER
OK OK ...you were CLOSE !!
let say the file is named "file1"
the simplest syntax that works is:
head -1|sed 's/ //g' > newfile
notice the single quotes in the sed command !
without them it returns a "sed: command garbled s/" error.
The syntax I used included cases for mulitple spaces or TABS. But since I am not using tabs YOU WIN !
Told ya it was easy !
let say the file is named "file1"
the simplest syntax that works is:
head -1|sed 's/ //g' > newfile
notice the single quotes in the sed command !
without them it returns a "sed: command garbled s/" error.
The syntax I used included cases for mulitple spaces or TABS. But since I am not using tabs YOU WIN !
Told ya it was easy !
ASKER
oh my solution ..sorry !
sed 's/^[ ]*//'
note: between the [ ] is a space and a tab character..the * means more than zero spaces or tabs ...basically ALL or them.
sed 's/^[ ]*//'
note: between the [ ] is a space and a tab character..the * means more than zero spaces or tabs ...basically ALL or them.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
ooops ...accepted the wrong answer ...sorry for those who pay to see it !
Have a good one.
Randy
Have a good one.
Randy
- Shell & Unix variant independent (awk isn't...)
Seems a short answer to your detailed spcification, but I think it meets all your requirements.