Edit certain variables in a file using bash script - bash

Question

Below i have the following SES.ses file:
1 VERSION_CONTROL {
2 FILE_NAME = "C:/Users/kjbaili/Documents/SWCD_TEST/SES.ses";
3 DATE = "";
4 VERSION = "1.1";
5 AUTHOR = "";
6 }
7
8 DISPLAYS {
9 DISPLAY xxx-c-vm-0120:main = {
10 COMPUTER = "xxx-C-VM-0120";
11 DISPLAY = "main";
12 ITEMS {
13 PANEL {
14 name = "visu.pnl";
15 }
16 }
17 }
18 }
19
20 RT-HOSTS {
21 RT-HOST xxx-c-vm-0120 = {
22 COMPONENT {
23 name = "RTE_connections_xxxxx.cmp";
24 }
25 COMPONENT {
26 name = "xxxx.cmp";
27 }
28 }
29 RT-HOST xxx-c-agx-0003 = {
30 COMPONENT {
31 name = "CtApxxx.cmp";
32 }
33 COMPONENT {
34 name = "CtApxxx.cmp";
35 }
36 COMPONENT {
37 name = "CtApxxx.cmp";
38 }
39 }
40 }
41
42 HARDWARE {
43 }
The user enters three Inputs to be placed in lines 2, 9, 21 and 29 -> /userpath, DISPLAY Node0, RT-HOSTS Node0, and RT-HOSTS Node1 accordingly.
I'm trying to edit variables in this file based on above user's input. These are : FILE_NAME (line 2), DISPLAY (line 9), RT-HOSTS ( line 21 and 29)
After doing some research i could build the following:
currentPath=$(pwd)/SES
awk -v path="$currentPath" '/FILE_NAME/ {cnt++} /FILE_NAME/ && cnt==1 {lnee=gensub(/(^.*\")(.*)(\".*$)/,"\\1"path"\\3",$0);print lnee; next}1' SES.ses > SES.temp && mv -f SES.tmp SES.ses
This command supposes to find the first entry for FILE_NAME and set the variable currentPath to it. However i'm getting the following error:
awk: cmd. line:1: warning: regexp escape sequence `\"' is not a known regexp operator
mv: cannot stat 'SES.tmp': No such file or directory
So my question is how to solve this error and how to set the other three variables line 9, 21 and 29
Thank in advance and would really appriciate your help
suggested solution from #Ed Morton:
awk -v filename='foo' -v display='bar' -v rthosts='some others' 'BEGIN { numRth = split(rthosts,rths) }
(filename != "") && ($1 == "FILE_NAME") { newval=filename; filename="" }
(display != "") && ($1 == "DISPLAY") { newval=display; display="" }
(numRth in rths) && ($1 == "RT-HOST") { newval=rths[++c]; delete rths[c] }
newval != "" {
oldval = ( $2 == "=" ? $3 : $2 )
gsub(/^[^"]*"|"[^"]*$/,"",oldval)
pos = index($0,oldval)
$0 = substr($0,1,pos-1) newval substr($0,pos+length(oldval))
newval = ""
}
{ print }' SES.ses
output of cat -Ev SES.ses
$
VERSION_CONTROL {$
FILE_NAME = "/c/Users/kjbaili/Documents/DO_NOT_DELETE/SES";$
DATE = "";$
VERSION = "1.1";$
AUTHOR = "";$
}$
$
DISPLAYS {$
DISPLAY d = {$
COMPUTER = "FDT-C-VM-0120";$
DISPLAY = "main";$
ITEMS {$
PANEL {$
name = "visu.pnl";$
}$
}$
}$
}$
$
RT-HOSTS {$
RT-HOST v = {$
COMPONENT {$
name = "RTE_connections_CtCoFallbackPath.cmp";$
}$
COMPONENT {$
name = "CtGwHwpFbpCmp.cmp";$
}$
}$
RT-HOST v = {$
COMPONENT {$
name = "CtApHwpFbpSit.cmp";$
}$
COMPONENT {$
name = "CtApHwpFbpMpl.cmp";$
}$
COMPONENT {$
name = "CtApHwpFbpCVGen.cmp";$
}$
}$
}$
$
HARDWARE {$
}$

Answer 1

The 3rd argument to the gensub() function must be a count of replacement
such as 1 or g (global).
Would you please try instead:
#!/bin/bash
# user's inputs
read -p "FILE_NAME: " -r file_name
read -p "DISPLAY: " -r display
read -p "RT-HOST: " -r rt_host
awk -v file_name="$file_name" -v display="$display" -v rt_host="$rt_host" '
{
sub(/FILE_NAME *= *\"[^"]+/, "FILE_NAME = \"" file_name)
sub(/DISPLAY *[^:]+:/, "DISPLAY " display ":")
}
/RT-HOST / {
if (! count++)
sub(/RT-HOST *[^[:space:]]+ *=/, "RT-HOST " display " =")
else
sub(/RT-HOST *[^[:space:]]+ *=/, "RT-HOST " rt_host " =")
}
1
' SES.ses > SES.tmp && mv -f -- SES.tmp SES.ses

Answer 2

You're getting the warning (not error) message escape sequence '\"' is not a known regexp operator because " isn't a regexp metachar, it's just a plain old literal char like x, but in your regexp (^.*\")(.*)(\".*$) you've written \" so either:
you're trying to escape " when it doesn't need to be escaped
or,
you're trying to include a literal \ in your regexp but
you'd need to escape it as \\ to do that.
So either way something is wrong with your regexp so awk is warning you about it.
I THINK this is probably what you're trying to do:
$ cat tst.awk
BEGIN { numRth = split(rthosts,rths) }
(filename != "") && ($1 == "FILE_NAME") { newval=filename; filename="" }
(display != "") && ($1 == "DISPLAY") { newval=display; display="" }
(numRth in rths) && ($1 == "RT-HOST") { newval=rths[++c]; delete rths[c] }
newval != "" {
oldval = ( $2 == "=" ? $3 : $2 )
gsub(/^[^"]*"|"[^"]*$/,"",oldval)
pos = index($0,oldval)
$0 = substr($0,1,pos-1) newval substr($0,pos+length(oldval))
newval = ""
}
{ print }
$ awk -v filename='foo' -v display='bar' -v rthosts='some others' -f tst.awk file
VERSION_CONTROL {
FILE_NAME = "foo";
DATE = "";
VERSION = "1.1";
AUTHOR = "";
}
DISPLAYS {
DISPLAY bar = {
COMPUTER = "xxx-C-VM-0120";
DISPLAY = "main";
ITEMS {
PANEL {
name = "visu.pnl";
}
}
}
}
RT-HOSTS {
RT-HOST some = {
COMPONENT {
name = "RTE_connections_xxxxx.cmp";
}
COMPONENT {
name = "xxxx.cmp";
}
}
RT-HOST others = {
COMPONENT {
name = "CtApxxx.cmp";
}
COMPONENT {
name = "CtApxxx.cmp";
}
COMPONENT {
name = "CtApxxx.cmp";
}
}
}
HARDWARE {
}

Answer 3

If ed is available/acceptable.
This should show the replacement.
#!/usr/bin/env bash
file_name=foobar
display=barmore
rt_host=quxfux
ed -s SES.ses <<-EOF
2,/FILE_NAME/s/".\{1,\}"/"$file_name"/
9s|^\([[:blank:]]*DISPLAY\).\{1,\}\(=[[:blank:]]*{\$\)|\1 $display \2|
21,29s|^\([[:blank:]]*RT-HOST\).\{1,\}\(=[[:blank:]]*{\$\)|\1 $rt_host \2|
,p
Q
EOF
Just like what #tshiono did using the builtin read to get the user's input
#!/usr/bin/env bash
# user's inputs
read -p "FILE_NAME: " file_name
read -p "DISPLAY: " display
read -p "RT-HOST: " rt_host
ed -s SES.ses <<-EOF
2,/FILE_NAME/s/".\{1,\}"/"$file_name"/
9s|^\([[:blank:]]*DISPLAY\).\{1,\}\(=[[:blank:]]*{\$\)|\1 $display \2|
21,29s|^\([[:blank:]]*RT-HOST\).\{1,\}\(=[[:blank:]]*{\$\)|\1 $rt_host \2|
,p
Q
EOF
The ,p is there just show what is the output of the newly edited buffer, remove it to silence the output.
Change Q to w to edit the file-inplace