I would like to run a GNU AWK script that is editing files in-place (e.g. with the -i inplace option) and have it print the filenames it is working on to STDOUT. Is there a way to do this? The following just adds the filename as the first line in the modified file, rather than printing the filename on the command line:
BEGINFILE {
print FILENAME
}
Here is a workaround; drop -i inplace from the command line (not an obligatory though, see -e/-f) and place following at the very beginning of your script. Before starting to process a file's content, this will disable inplace temporarily and print FILENAME. Then inplace's BEGINFILE rule will enable itself again.
BEGINFILE {
if (inplace::filename != "") {
inplace::end(inplace::filename, inplace::suffix)
inplace::filename = ""
}
print FILENAME
}
#include "inplace"
See how inplace is implemented for a better understanding.
Related
I have a script that looks like this:
#! /bin/awk -f
BEGIN { print "start" }
{ print $0 }
END { print "end" }
Call the script like this: ./myscript.awk test.txt
Pretty simple - takes a file and adds "start" to the start and "end" to the end.
Now I want to take the input filename, lets call it test.txt, and print the output to a file called test.out.
So I tried to print the input filename:
BEGIN { print "fname: '" FILENAME "'" }
But that printed: fname: '' :(
The rest I can figure out I think, I have this following to print to a hard-coded filename:
#! /bin/awk -f
BEGIN { print "start" > "test.out" }
{ print $0 >> "test.out" }
END { print "end" >> "test.out" }
And that works great.
So the questions are:
how do I get the input filename?
Assuming somehow I get the input file name in a variable, e.g. FILENAME which contains "test.txt" how would I make another variable, e.g. OUTFILE, which contains "test.out"?
Note: I will be doing much more awk processing so please don't suggest to use sed or other languages :))
Try something like this:
#! /bin/awk -f
BEGIN {
file = gensub(".txt",".out","g",ARGV[1])
print "start" > file
}
{ print $0 >> file }
END {
print "end" >> file
close(file)
}
I'd suggest to close() the file too in the END{} statement. Good call to Sundeep for pointing out that FILENAME is empty in BEGIN.
$ echo 'foo' > ip.txt
$ awk 'NR==1{op=FILENAME; sub(/\.[^.]+$/, ".log", op); print "start" > op}
{print > op}
END{print "end" > op}' ip.txt
$ cat ip.log
start
foo
end
Save FILENAME to a variable, change the extension using sub and then print as required
From gawk manual
Inside a BEGIN rule, the value of FILENAME is "", because there are no input files being processed yet
If you're using GNU awk (gawk), you can use the patterns BEGINFILE and ENDFILE
awk 'BEGINFILE{
outfile=FILENAME;
sub(".txt",".out",outfile);
print "start" > outfile
}
ENDFILE{
print "stop" >outfile
}' file1.txt file2.txt
You can then use the variable outfile your the main {...} loop.
Doing so will allow you to process more that 1 file in a single awk command.
I have an smb.conf ini file which is overwritten whenever edited with a certain GUI tool, wiping out a custom setting. This means I need a cron job to ensure that one particular section in the file contains a certain option=value pair, and insert it at the end of the section if it doesn't exist.
Example
Ensure that hosts deny=192.168.23. exists within the [myshare] section:
[global]
printcap name = cups
winbind enum groups = yes
security = user
[myshare]
path=/mnt/myshare
browseable=yes
enable recycle bin=no
writeable=yes
hosts deny=192.168.23.
[Another Share]
invalid users=nobody,nobody
valid users=nobody,nobody
path=/mnt/share2
browseable=no
Long-winded solution using awk
After a long time struggling with sed, I concluded that it might not be the right tool for the job. So I moved over to awk and came up with this:
#!/bin/sh
file="smb.conf"
tmp="smb.conf.tmp"
section="myshare"
opt="hosts deny=192.168.23."
awk '
BEGIN {
this_section=0;
opt_found=0;
}
# Match the line where our section begins
/^[ \t]*\['"$section"'\][ \t]*$/ {
this_section=1;
print $0;
next;
}
# Match lines containing our option
this_section == 1 && /^[ \t]*'"$opt"'[ \t]*$/ {
opt_found=1;
}
# Match the following section heading
this_section == 1 && /^[ \t]*\[.*$/ {
this_section=0;
if (opt_found != 1) {
print "\t'"$opt"'";
}
}
# Print every line
{ print $0; }
END {
# In case our section is the very last in the file
if (this_section == 1 && opt_found != 1) {
print "\t'"$opt"'";
}
}
' $file > $tmp
# Overwrite $file only if $tmp is different
diff -q $file $tmp > /dev/null 2>&1
if [ $? -ne 0 ]; then
mv $tmp $file
# reload smb.conf here
else
rm $tmp
fi
I can't help feeling that this is a long script to achieve a simple task. Is there a more efficient/elegant way to insert a property in an ini file using basic shell tools like sed and awk?
Consider using Python 3's configparser:
#!/usr/bin/python3
import sys
from configparser import SafeConfigParser
cfg = SafeConfigParser()
cfg.read(sys.argv[1])
cfg['myshare']['hosts deny'] = '192.168.23.';
with open(sys.argv[1], 'w') as f:
cfg.write(f)
To be called as ./filename.py smb.conf (i.e., the first parameter is the file to change).
Note that comments are not preserved by this. However, since a GUI overwrites the config and doesn't preserve custom options, I suspect that comments are already nuked and that this is not a worry in your case.
Untested, should work though
awk -vT="hosts deny=192.168.23" 'x&&$0~T{x=0}x&&/^ *\[[^]]+\]/{print "\t\t"T;x=0}
/^ *\[myshare\]/{x++}1' file
This solution is a bit awkward. It uses the INI section header as the record separator. This means that there is an empty record before the first header, so when we match the header we're interested in, we have to read the next record to handle that INI section. Also, there are some printf commands because the records still contain leading and trailing newlines.
awk -v RS='[[][^]]+[]]' -v str="hosts deny=192.168.23." '
{printf "%s", $0; printf "%s", RT}
RT == "[myshare]" {
getline
printf "%s", $0
if (index($0, str) == 0) print str
printf "%s", RT
}
' smb.conf
RS is the awk variable that contains the regex to split the text into records.
RT is the awk variable that contains the actual text of the current record separator.
With GNU awk for a couple of extensions:
$ cat tst.awk
index($0,str) { found = 1 }
match($0,/^\s*\[([^]]+).*/,a) {
if ( (name == tgt) && !found ) { print indent str }
name = a[1]
found = 0
}
{ print; indent=gensub(/\S.*/,"","") }
.
$ awk -v tgt="myshare" -v str="hosts deny=192.168.23." -f tst.awk file
[global]
printcap name = cups
winbind enum groups = yes
security = user
[myshare]
path=/mnt/myshare
browseable=yes
enable recycle bin=no
writeable=yes
hosts deny=192.168.23.
[Another Share]
invalid users=nobody,nobody
valid users=nobody,nobody
path=/mnt/share2
browseable=no
.
$ awk -v tgt="myshare" -v str="fluffy bunny" -f tst.awk file
[global]
printcap name = cups
winbind enum groups = yes
security = user
[myshare]
path=/mnt/myshare
browseable=yes
enable recycle bin=no
writeable=yes
hosts deny=192.168.23.
fluffy bunny
[Another Share]
invalid users=nobody,nobody
valid users=nobody,nobody
path=/mnt/share2
browseable=no
I am writing a report tool which processes the source files of some application and produce a report table with two columns, one containing the name of the file and the other containing the word TODO if the file contains a call to some deprecated function deprecated_function and DONE otherwise.
I used awk to prepare this report and my shell script looks like
report()
{
find . -type f -name '*.c' \
| xargs -n 1 awk -v deprecated="$1" '
BEGIN { status = "DONE" }
$0 ~ deprecated{ status = "TODO" }
END {
printf("%s|%s\n", FILENAME, status)
}'
}
report "deprecated_function"
The output of this script looks like
./plop-plop.c|DONE
./fizz-boum.c|TODO
This works well but I would like to rewrite the awk script so that it supports several input files instead of just one — so that I can remove the -n 1 argument to xargs. The only solutions I could figure out involve a lot of bookkeeping, because we need to track the changes of FILENAME and the END event to catch each end of file event.
awk -v deprecated="$1" '
BEGIN { status = "DONE" }
oldfilename && (oldfilename != FILENAME) {
printf("%s|%s\n", oldfilename, status);
status = DONE;
oldfilename = FILENAME;
}
$0 ~ deprecated{ status = "TODO" }
END {
printf("%s|%s\n", FILENAME, status)
}'
Maybe there is a cleaner and shorter way to handle this.
I am using FreeBSD's awk and am looking for solutions compatible with this tool.
This will work in any modern awk:
awk -v deprecated="$1" -v OFS='|' '
$0 ~ deprecated{ dep[FILENAME] }
END {
for (i=1;i<ARGC;i++)
print ARGV[i], (ARGV[i] in dep ? "TODO" : "DONE")
}
' file1 file2 ...
Any time you need to produce a report for all files and don't have GNU awk for ENDFILE, you MUST loop through ARGV[] in the END section (or loop through it in BEGIN and populate a different array for END section processing). Anything else will fail if you have empty files.
Your awk script could be something like this:
awk -v deprecated="$1" '
FNR==1 {if(file) print file "|" (f?"TODO":"DONE"); file=FILENAME; f=0}
$0 ~ deprecated {f=1}
END {print file "|" (f?"TODO":"DONE")}' file1.c file2.c # etc.
The logic is fairly similar to your program so hopefully it's all clear. FNR is the record number of the current file, which I'm using to detect the start of a new file. Admittedly there's some repetition in the END block but I don't think it's a big deal. You could always use a function if you wanted to.
Testing it out:
$ cat f1.c
int deprecated_function()
{
// some deprecated stuff
}
$ cat f2.c
int good_function()
{
// some good stuff
}
$ find -name "f?.c" -print0 | xargs -0 awk -v deprecated="deprecated" 'FNR==1 {if(file) print file "|" (f?"TODO":"DONE"); file=FILENAME; f=0} $0 ~ deprecated {f=1} END {print file "|" (f?"TODO":"DONE")}'
./f2.c|DONE
./f1.c|TODO
I have used -print0 and the -0 switch to xargs so that both programs with work file names separated by null bytes "\0" rather than spaces. This means that you won't run into problems with spaces in file names.
I am trying to append lines to some new files with awk in this way:
#!/usr/bin/awk -f
BEGIN {
FS = "[ \t|]"; }
{
print $5 "\t" $13 "\t" $14 >> "./bed/" $5 ".bed";
}
END {
}
New file is created with filename derived from a field of awk input file (5th field). I am unable to execute this script since it fails with
awk: ./blast2bed.awk:6: (FILENAME=blastout000 FNR=1) fatal: can't redirect to `./bed/AY517392.1.bed' (No such file or directory)
Any hints?
Thanks
The directory bed has to exist so create it first with mkdir bed either before you run your script or in the BEGIN block. You should also add brackets around the output file:
print $5"\t"$13"\t"$14 >> ("./bed/"$5".bed")
Notes: You don't need to end lines with ; if you have a single statement per line and the BEGIN and END blocks are optional.
I'm trying to run the command below, and its giving me the error. Thoughts on how to fix? I would rather have this be a one line command than a script.
grep "id\": \"http://room.event.assist.com/event/room/event/" failed_events.txt |
head -n1217 |
awk -F/ ' { print $7 } ' |
awk -F\" ' { print "url \= \"http\:\/\/room\.event\.assist\.com\/event\/room\/event\/'{ print $1 }'\?schema\=1\.3\.0\&form\=json\&pretty\=true\&token\=582EVTY78-03iBkTAf0JAhwOBx\&account\=room_event\"" } '
awk: non-terminated string url = "ht... at source line 1
context is
>>> <<<
awk: giving up
source line number 2
The line below exports out a single column of ID's:
grep "id\": \"http://room.event.assist.com/event/room/event/" failed_events.txt |
head -n1217 |
awk -F/ ' { print $7 } '
156512145
898545774
454658748
898432413
I'm looking to get the ID's above into a string like so:
" url = "string...'ID'string"
take a look what you have in last awk :
awk -F\"
' #single start here
{ print " #double starts for print, no ends
url \= \"http\:\/\/room\.event\.assist\.com\/event\/room\/event\/
' #single ends here???
{ print $1 }'..... #single again??? ...
(rest codes)
and you want to print exact {print } out? i don't think so. why you were nesting print ?
Most of the elements of your pipe can be expressed right inside awk.
I can't tell exactly what you want to do with the last awk script, but here are some points:
Your "grep" is really just looking for a string of text, not a
regexp.
You can save time and simplify things if you use awk's
index() function instead of a RE. Output formats are almost always
best handled using printf().
Since you haven't provided your input data, I can't test this code, so you'll need to adapt it if it doesn't work. But here goes:
awk -F/ '
BEGIN {
string="id\": \"http://room.event.assist.com/event/room/event/";
fmt="url = http://example.com/event/room/event/%s?schema=whatever\n";
}
count == 1217 { nextfile; }
index($0, string) {
split($7, a, "\"");
printf(fmt, a[0]);
count++;
}' failed_events.txt
If you like, you can use awk's -v option to pass in the string variable from a shell script calling this awk script. Or if this is a stand-alone awk script (using #! shebang), you could refer to command line options with ARGV.