sed/awk/perl formatting multiple paragraphs
Question
dzinfo gives me output formatted like this;
User: x0000001
Forced into restricted environment: No
Role Name Avail Restricted Env
--------------- ----- --------------
login/Corp_All Yes None
_Example-Role-- Yes _Example-Role--
ALL_Servers-Win ALL_Servers-Win
/US_All /US_All
Domain_GLOBAL-Ro Yes Domain_GLOBAL-Ro
le-CORE_Group-AL le-CORE_Group-AL
L-MacOS/Domain_ L-MacOS/Domain_
GLOBAL GLOBAL
Effective rights:
Password login
Non password login
Allow normal shell
PAM Application Avail Source Roles
--------------- ----- --------------------
* Yes login/US_All
Privileged commands:
Name Avail Command Source Roles
--------------- ----- -------------------- --------------------
CORP_GLOBAL-Com Yes /usr/bin/getfacl CORP_GLOBAL-Role-COR
mand-CORE_SVR_I E_SVR_INFRA_ALL-LNX/
NFRA_ALL-V042-S CORP_GLOBAL
00042/CORP_GLOB
AL
CORP_GLOBAL-Com Yes /usr/bin/dzdo -l CORP_GLOBAL-Role-COR
mand-CORE_SVR_I E_SVR_INFRA_ALL-LNX/
NFRA_ALL-V042-S CORP_GLOBAL
00048/CORP_GLOB
AL
CORP_GLOBAL-Com Yes /bin/cp temp_auth CORP_GLOBAL-Role-COR
mand-CORE_SVR_I /home/sudocfg/author E_SVR_INFRA_ALL-LNX/
NFRA_ALL-V042-S ized_keys CORP_GLOBAL
00085/CORP_GLOB
AL
What tool would be the best choice to format a report like this? And how could I match/combine and format the columns/lines to something like the following?
User: x0000001
Forced into restricted environment: No
Role Name Avail Restricted Env
--------------- ----- --------------
login/Corp_All Yes None
_Example-Role--ALL_Servers-Win/US_All Yes _Example-Role--ALL_Servers-Win/US_All
Domain_GLOBAL-Role-CORE_Group-ALL-MacOS/Domain_GLOBAL Yes Domain_GLOBAL-Role-CORE_Group-ALL-MacOS/Domain_GLOBAL
Effective rights:
Password login
Non password login
Allow normal shell
PAM Application Avail Source Roles
--------------- ----- --------------------
* Yes login/US_All
Privileged commands:
Name Avail Command Source Roles
--------------- ----- -------------------- --------------------
CORP_GLOBAL-Command-CORE_SVR_INFRA_ALL-V042-S00042/CORP_GLOBAL Yes /usr/bin/getfacl CORP_GLOBAL-Role-CORE_SVR_INFRA_ALL-LNX/CORP_GLOBAL
CORP_GLOBAL-Command-CORE_SVR_INFRA_ALL-V042-S00048/CORP_GLOBAL Yes /usr/bin/dzdo -l CORP_GLOBAL-Role-CORE_SVR_INFRA_ALL-LNX/CORP_GLOBAL
CORP_GLOBAL-Command-CORE_SVR_INFRA_ALL-V042-S00085/CORP_GLOBAL Yes /bin/cp temp_auth /home/sudocfg/authorized_keys CORP_GLOBAL-Role-CORE_SVR_INFRA_ALL-LNX/CORP_GLOBAL
The text in each column can vary greatly, so I'd like to have the width automatically adjust.
I can handle one-liners, but for a report like this? I wouldn't know where to even begin.
2 answers
solution1
0 2016-12-15 20:12:07
To get you started, here's basically how you'd start to isolate the individual fields on each line using GNU awk for FIELDWIDTHS:
$ cat tst.awk
BEGIN { origFS=FS }
/---/ {
origFS=FS
split($0,f,/\s+|-+/,s)
FIELDWIDTHS=""
for (i=1; i in s; i++) {
FIELDWIDTHS = (i>1 ? FIELDWIDTHS " " : "") length(s[i])
}
}
/^\s*$/ {
FIELDWIDTHS=""
FS=origFS
}
{
for (i=1; i<=NF; i++) {
printf "<%s>", $i, (i<NF?OFS:ORS)
}
print ""
}
.
$ awk -f tst.awk file
<User:><x0000001>
<Forced><into><restricted><environment:><No>
<Role><Name><Avail><Restricted><Env>
<---------------><-----><-------------->
< ><login/Corp_All >< ><Yes >< ><None >< >
< ><_Example-Role-->< ><Yes >< ><_Example-Role-><->
< ><ALL_Servers-Win>< >< >< ><ALL_Servers-Wi><n>
< ></US_All >< >< >< ></US_All ><>
< ><Domain_GLOBAL-R><o ><Yes >< ><Domain_GLOBAL-><R>
< ><le-CORE_Group-A><L >< >< ><le-CORE_Group-><A>
< ><L-MacOS/Domain_>< >< >< ><L-MacOS/Domain><_>
< ><GLOBAL >< >< >< ><GLOBAL ><>
<Effective><rights:>
<Password><login>
<Non><password><login>
<Allow><normal><shell>
<PAM><Application><Avail><Source><Roles>
<---------------><-----><-------------------->
<* >< ><Yes >< ><login/US_All >< >
<Privileged comm><an><ds:><><><>
< Name >< >< Ava><i><l Command >< >
< -------------><-->< ---><-><- ------------------><->
< ><CORP_GLOBAL-Com>< ><Yes >< ></usr/bin/getfacl >< ><CORP_GLOBAL-Role-COR>< >
< ><mand-CORE_SVR_I>< >< >< >< >< ><E_SVR_INFRA_ALL-LNX/>< >
< ><NFRA_ALL-V042-S>< >< >< >< >< ><CORP_GLOBAL >< >
< ><00042/CORP_GLOB>< >< >< >< >< >< >< >
< ><AL >< >< >< >< >< >< >< >
< ><CORP_GLOBAL-Com>< ><Yes >< ></usr/bin/dzdo -l >< ><CORP_GLOBAL-Role-COR>< >
< ><mand-CORE_SVR_I>< >< >< >< >< ><E_SVR_INFRA_ALL-LNX/>< >
< ><NFRA_ALL-V042-S>< >< >< >< >< ><CORP_GLOBAL >< >
< ><00048/CORP_GLOB>< >< >< >< >< >< >< >
< ><AL >< >< >< >< >< >< >< >
< ><CORP_GLOBAL-Com>< ><Yes >< ></bin/cp temp_auth >< ><CORP_GLOBAL-Role-COR>< >
< ><mand-CORE_SVR_I>< >< >< ></home/sudocfg/author>< ><E_SVR_INFRA_ALL-LNX/>< >
< ><NFRA_ALL-V042-S>< >< >< ><ized_keys >< ><CORP_GLOBAL >< >
< ><00085/CORP_GLOB>< >< >< >< >< >< >< >
< ><AL >< >< >< >< ><><><>
You're going to want to buy the book Effective Awk Programming, 4th Edition, by Arnold Robbins and have it handy for reference as you start tinkering with that to see how it works and then building on it.
solution2
0 2018-10-24 07:56:29
I can handle one-liners, but for a report like this? I wouldn't know where to even begin.
Indeed this seems hardly possible with a one-liner (unless it's one very long line). But to begin with, the problem can be analyzed and a possible solution outlined.
- Identify what parts of the input constitute the tables (where column lines are to be combined): We could use the lines of hyphens between column headers and text to identify the start of a table, and a short line or an unfitting line, where the gaps between the columns are not blank, to identify the end.
- Read a table in line by line, split each line into the columns (whose margins can be derived from the hyphen lines), and, if the current line is a wrapped continuation line of the preceding line (ie if it has empty columns), concatenate each column's texts. In the course of this, the required column widths can be determined from the maximum combined text lengths.
- Write the table out, using the before determined column widths.
The following Perl program implements a solution. Note however that, since it uses seek , it cannot work on piped input, but only on a file.
#!/usr/bin/perl -wn
if (/---/) # here comes a table
{
seek ARGV, $priprior, 0; # set position back to header line
push @gap, $-[1] while /(?<=-) *( )-/g; # find the column gaps
$tabrow = -1; # index of current table row: no rows yet
@maxlng = (0)x($#gap+2); # maximum length of text in each column
while (<>) # read the table, starting with headers
{ # check for end of table:
last if length() <= $gap[-1] or substr($_, $gap[-1], 1) ne ' ';
$offset = 0; # first column start
$contin = 0; # is it a continuation line?
foreach $i (0..$#gap+1)
{ # extract column:
if ($i <= $#gap) # not last column?
{ $column[$i] = substr $_, $offset, $gap[$i]-$offset;
$offset = $gap[$i];
}
else # last column
{ $column[$i] = substr $_, $offset;
}
$column[$i] =~ s/\s+$//; # remove trailing whitespace
$contin += $column[$i] eq '' # column empty?
}
++$tabrow unless $contin;
foreach $i (0..$#gap+1)
{ # ugly fix to restore the space in "temp_auth /home/...", where
# the column is wrapped early: --------------------
# ...
# /bin/cp temp_auth
# /home/sudocfg/author
# ized_keys
$table[$tabrow][$i] .= ' ' if $contin and length($table[1][$i])>
length($table[$tabrow][$i]);
# now that's fixed, proceed with normal unwrapping of the column
$column[$i] =~ s/^ +// if $contin or $i; # remove leading spaces
$table[$tabrow][$i] .= $column[$i];
$maxlng[$i] = length $table[$tabrow][$i] # update maximum length
if $maxlng[$i] < length $table[$tabrow][$i];
}
undef $_; last if eof
}
foreach $e (@table)
{
foreach $i (0..$#gap+1) { printf "%-*s", $maxlng[$i]+1, $$e[$i]; }
print "\n";
}
undef @gap; undef @table;
}
else { print $previous_line if $previous_line }
$previous_line = $_;
$priprior = $prior;
$prior = tell ARGV;
END { print $previous_line if $previous_line }
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.
Related Question
Formatting Text Awk Sed
Sed or Awk or Perl substitution in a sentence
Specific data formatting using awk or sed
Sed, awk or perl to filter incomplete data column
compare the length of multiple files using awk or sed
match regexp against variable string in perl/sed/awk
Extract domain then paste into the same line using sed/awk/grep/perl
Extract all columns which match a string awk sed python perl
piping sed | awk or awk | sed
Extract columns from multiple text files with bash or awk or sed?
Related Tags