|
|
John Meissner
|
|
May 20, 2003 14:31:19 GMT
|
|
I am trying to replace and entire address in one file with the entire address in another file. The following is an excerpt from the first file (called core11.cfg):
# ftp and remsh sources can use a full path:
# archive_path = "/pub/IUXarchives/B.11.00_32bitCDE.gz"
# The data for the "impacts" statements are found
# by running /opt/ignite/lbin/archive_impact
impacts = "/" 27Kb
impacts = "/.dt" 35Kb
impacts = "/etc" 1864Kb
impacts = "/export" 1Kb
impacts = "/home" 1Kb
impacts = "/opt" 74096Kb
impacts = "/sbin" 13449Kb
impacts = "/stand" 1Kb
impacts = "/tmp" 1Kb
impacts = "/users" 40Kb
impacts = "/usr" 225951Kb
impacts = "/var" 5705Kb
exrequisite += "golden image - 64 bit OS"
visible_if = can_run_32bit
and this is the entire 2nd file (called mou110.impact):
impacts = "/" 23Kb
impacts = "/.sw" 100Kb
impacts = "/dev" 14Kb
impacts = "/etc" 41366Kb
impacts = "/home" 8Kb
impacts = "/opt" 1246434Kb
impacts = "/sbin" 37158Kb
impacts = "/stand" 55954Kb
impacts = "/tmp" 6Kb
impacts = "/users" 1Kb
impacts = "/usr" 909819Kb
impacts = "/var" 542014Kb
I'm trying to replace all the "impact" entried in the first file with the entried in the second file. The reasone I am trying to replace the whole section is that the entried are not all the same....
i.e.
/.dt (original file)
/.sw (replacement file)
I'm in the process of scripting the creation of a bootable ingnite DVD (iso image)
Any help would be appreciated.
Thanks
|
|
|
Note: If you are the author of this question and wish to assign points to any of the answers, please login first.For more information on assigning points ,click
here
|
|
|
Sort Answers By:
Date or Points
|
|
|
Ian Dennison
|
|
May 20, 2003 14:40:48 GMT
3 pts
|
|
John,
Replace all entries named "impacts" in the first file with contents of the second file? Is that the right interpretation?
Ian |
|
Rodney Hills
|
|
May 20, 2003 14:43:53 GMT
5 pts
|
|
If we assume the comment line before the first "impacts" stays the same-
sed -e '/^impacts/d' -e '/^# by running/r mou110.impact' <core11.cfg >newcore11.cfg
This will delete all "impacts" line from original and "r"ead the replacment file following the comment.
HTH
-- Rod Hills |
|
|
John Meissner
|
|
May 20, 2003 14:44:32 GMT
N/A: Question Author
|
|
|
sorry - I should clarify.... i just want to replace the lines that start with "impact" |
|
|
Ian Dennison
|
|
May 20, 2003 14:51:31 GMT
10 pts
|
|
rm /tmp/newfile
cat /tmp/file1 |awk 'BEGIN {indic=0} $1 ~ /impact/ && indic == 0 {indic=1;rc=system("cat /tmp/impacts")}
indic == 1 && $1 !~ /impact/ {indic=2}
indic == 0 && $1 !~ /impact/ {print $0}
indic == 2 {print $0} ' >/tmp/newfile
Sample code, testing and works. /tmp/newfile = new output, /tmp/impacts = new list of impacts (no blank lines).
Share and Enjoy! Ian |
|
|
Massimo Bianchi
|
|
May 20, 2003 14:51:33 GMT
2 pts
|
|
I would suggest a clean way:
#!/sbin/sh
TMP=$$
cat core11.cfg | grep -v impact > /tmp/$TMP.spool
cat /tmp/$TMP.spool > core11.cfg
cat mou110.impact >> core11.cfg
####EOF
Doing so you will have all the variable where you want.
HTH,
Massimo |
|
|
John Meissner
|
|
May 20, 2003 14:52:02 GMT
N/A: Question Author
|
|
|
rodney - yes the previous commented line will stay the same. I tried your line but it didn't change anything in the file. |
|
|
Steve Steel
|
|
May 20, 2003 14:57:45 GMT
2 pts
|
|
Hi
using file1 file2 in,temp as work and file3 out
grep -v impacts file1 > temp
grep "#" temp > file3
cat file2 >> file3
grep -v "#" temp >> file3
rm temp
Steve Steel |
|
|
Rodney Hills
|
|
May 20, 2003 15:01:18 GMT
5 pts
|
|
I put your 2 files on my system and run the following-
sed -e '/^impacts/d' -e '/^# by running/r testf2' <testf1
and this is the results-
# ftp and remsh sources can use a full path:
# archive_path = "/pub/IUXarchives/B.11.00_32bitCDE.gz"
# The data for the "impacts" statements are found
# by running /opt/ignite/lbin/archive_impact
impacts = "/" 23Kb
impacts = "/.sw" 100Kb
impacts = "/dev" 14Kb
impacts = "/etc" 41366Kb
impacts = "/home" 8Kb
impacts = "/opt" 1246434Kb
impacts = "/sbin" 37158Kb
impacts = "/stand" 55954Kb
impacts = "/tmp" 6Kb
impacts = "/users" 1Kb
impacts = "/usr" 909819Kb
impacts = "/var" 542014Kb
exrequisite += "golden image - 64 bit OS"
visible_if = can_run_32bit
I had both testf1 and testf2 both under my current directory.
-- Rod Hills |
|
|
Rodney Hills
|
|
May 20, 2003 15:04:50 GMT
1 pts
|
|
Remember, "sed" does not change the original file. It creates a "new" text file.
If you want to replace the original, then "mv" the resultant file to the original.
-- Rod Hills |
|
|
John Meissner
|
|
May 20, 2003 15:05:28 GMT
N/A: Question Author
|
|
|
Ian - your solution does indeed work. Just so everyone else knows... the original file is much larger than I included in my first post. I'm attaching the entire file just so there is no misunderstandings. |
|
|
Ian Dennison
|
|
May 20, 2003 15:08:46 GMT
3 pts
|
|
John,
Have just noticed the recursion within your sample file (2 sets of "impacts" lines).
Do we need to code around this, by either
a) replacing both sets of "impacts" lines, or
b) only replacing a specific set of "impacts" lines?
Have solutions to both; let us know if you need them, and which one.
Share and Enjoy! Ian |
|
|
Curtis Larson
|
|
May 20, 2003 15:12:12 GMT
5 pts
|
|
I'd do something this:
#!/usr/bin/ksh
file1=
file2=
tmpFile=
rm $tmpFile
cat file1 |
while read a b c size
do
start="$a $b $c"
case $start in
^${start}*)
size=$(grep "^$start" file2 | cut -f4)
;;
esac
print "$start $size" >> $tmpFile
done
haven't tested but should be pretty close to what your looking to do |
|
|
John Meissner
|
|
May 20, 2003 15:12:33 GMT
N/A: Question Author
|
|
|
rodney - I'm outputting your script into a new file and it didn't change anything. not sure why. |
|
|
Jean-Louis Phelix
|
|
May 20, 2003 15:13:50 GMT
5 pts
|
|
hi,
This awk script could also do it ...
Regards.
awk '/^impact/ {if (flg != 1)
{
flg=1
system("cat mou110.impact")
}
next
}
{print}' core11.cfg |
|
|
John Meissner
|
|
May 20, 2003 15:15:18 GMT
N/A: Question Author
|
|
|
Ian - both sets would be great - I just wasn't sure how to do the entire addressing issue. |
|
|
Ian Dennison
|
|
May 20, 2003 15:34:41 GMT
10 pts
|
|
# Second code - Checks for keywords that indicate we have found our subsection (Reset every time we hit a right curly brace)
cat /tmp/file1 |awk -vsubset="64 bit" 'BEGIN {indic=0;subsect=0}
$0 ~ "\}" {indic=0;subsect=0}
$0 ~ subset {subsect=1}
$1 ~ /impact/ && indic == 0 && subsect == 1 {indic=1;rc=system("cat /tmp/impacts")}
indic == 1 && $1 !~ /impact/ {indic=2}
indic == 0 && $1 ~ /impact/ && subsect != 1 {print $0}
indic == 0 && $1 !~ /impact/ {print $0}
indic == 2 {print $0} ' >/tmp/newfile
# First code - resets back to Zero everytime if finds a curly brace (does all occurences)
# cat /tmp/file1 |awk 'BEGIN {indic=0}
# $0 ~ "\{" || $0 ~ "\}" {indic=0}
# $1 ~ /impact/ && indic == 0 {indic=1;rc=system("cat /tmp/impacts")}
# indic == 1 && $1 !~ /impact/ {indic=2}
# indic == 0 && $1 !~ /impact/ {print $0}
# indic == 2 {print $0} ' >/tmp/newfile
Attached as a file. First piece of code is commented out so I could test the second. Both pieces tested on your sample, second part tested on keywords "32 bit" and "64 bit".
Share and Enjoy! Ian |
|
|
John Meissner
|
|
May 20, 2003 15:46:35 GMT
N/A: Question Author
|
|
|
Ian - your script does exactly what I was looking for. I'm still testing some of the others and will report back and assign points to the rest. |
|
|
John Meissner
|
|
Jun 2, 2003 17:06:06 GMT
N/A: Question Author
|
|
Ian -
It's been a while since you solved this problem... I was bored today and was looking over this ... I've never used this type of structure (indic) before and was wondering if you could explain it a little?
if not thanks for the solution....
catch a man a fish and you feed him for a day.... teach a man to fish and you feed him for a lifetime. |
|
Printable version
