最后用 sed 和 awk 解析以创建 csv 文件

Question

So, I am trying to create a awk script to go through the output of the last command so that I can convert it to a csv file. The problem I am having is that lines are matching several times and I am not sure how to prevent it from happening; and I am by no means an awk expert ( part of the reason for this exercise )

The reason for the sed insertions are for missing "columns".

command to massage last command and invoke the awk script:

last -w -F | sed -r 's/[[:space:]{23,}/ unknown /;s/(crash|down)/\1 unknown /' | awk -f awktest.awk

( as a side note, the sub in here was before I decided to test inserting columns with sed )

awktest.awk

BEGIN { print "user,tty/pts,connection_from,login_time,state,logoff_time,total_time"}
{if ($1 ~/reboot/) {print $1","$2","$3","$4","($5" "$6" "$7" "$8" "$9)","$10","($11" "$12" "$13" "$14" "$15","$16)}}
{if ($1 ~/root/ && $2 ~/tty/) { sub($3,"null") ;  print $1","$2","$3","($4" "$5" "$6" "$7" "$8)","$9","($10","$11","$12" "$13" "$14)}}

#{if ( NF > 14) { print $1","$2","$3","($4" "$5" "$6" "$7" "$8 )","$9","($10" "$11" "$12" "$13" "$14)","$15 }}
{if ($3 == "unknown" && $11 !="unknown") { print $1","$2","$3","($4" "$5" "$6" "$7" "$8 )","$9","($10" "$11" "$12" "$13" "$14)","$15 }}
#{if ($3 == "unknown") { print $1","$2","$3","( $4" "$5" "$6" "$7" "$8 )","$9","($10","$11",")$12}}
{if ($11 == "unknown") { print $1","$2","$3","( $4" "$5" "$6" "$7" "$8 )","$9","($10","$11",")$12}}

As you can see, I have tried multiple ways , and there are either duplicates or there are missing lines. The duplicates are clear:

root,tty2,null,Sun Jul 19 13:25:38 2020,-,down,unknown,(00:01)
root,tty2,null,Sun Jul 19 13:25:38 2020,-,down,unknown,(00:01)

Yeah, I know, not pretty; but I was experimenting; and I am stuck and have been reading the sed and awk book and googling like crazy and making no headway; and at this point I think I have looked too long at it to make any progress.

What am I doing wrong?

# edit for clarity

( some whitespace was removed to make it cleaner) Examples of last command :

root  tty3         Mon Jun  8 09:49:56 2020 - down           (00:00)
foo :0    0        Mon Jun  8 09:49:16 2020 - down                      (00:01)
reboot   system boot  5.6.16-300.fc32.x86_64 Mon Jun  8 09:48:28 2020 - Mon Jun  8 09:50:54 2020  (00:02)
roncioiu :0          Thu Jun 18 10:19:29 2020 - Thu Jun 18 10:20:19 2020  (00:00)

# desired state columns:

• the time stamps count as one column as they are in ()

"user,tty/pts,connection_from,login_time,state,logoff_time,total_time"

Answer 1

So I eventually settled on a one liner and abandoned the script thanks to gentle prodding by @tink . With just a little bit of data munging.

Thanks for the tip! I was losing my sanity doing it the other way.

last -a -w --time-format iso | sed -r 's/\sboot\s//g;/^$/d;/wtmp/d;/(crash|down)/s/$/ unknown /' | awk 'BEGIN { print "user,pts/tty,login_time,state,logoff_time,total_time,host "} { print $1","$2","$3","$4","$5","$6","$7}'