问题
I had amazing help on an AWK script here and thought to myself it would be really cool to have the exact same output I am monitoring on the CLI to go to a CSV file. I did research and found a great answer here, it basically showed code like this:
awk '{print $1","$2","$3","$4","$5}' < /tmp/file.txt > /tmp/file.csv
The first issue I have is /tmp/file.txt is not needed as my code is already producing the string with separated values. I don't know if my variables would work without running all new AWK commands, so I would prefer to just tag it to the end of the previous AWK command if possible. But I don't know how to implement the same concept within the actual script I am using. Could anyone show me the formatting schema I would need to tag this into the end of my script?
My ever-evolving script looks like this:
#!/bin/bash
CURRENT_DATE=`date +%Y-%m-%d`
tail -fn0 /var/log/pi-star/MMDVM-"$CURRENT_DATE".log | gawk '
match($0, /received.*voice header from ([[:alnum:]]+) to ([[:alnum:]]+
[0-9]+)/, a) {
in_record = 1
call_sign = a[1]
channel = a[2]
}
in_record && match($0, /DMR ID: ([0-9]+)/, a) {
dmr_id = a[1]
}
in_record && match($0, /([0-9.]+) seconds, ([0-9]+)% packet loss, BER:
([0-9.]+)%/, a) {
in_record = 0
print call_sign, channel, dmr_id, a[1], a[2], a[3]
}
' OFS=,
done
I still want to monitor via the terminal, I just think the appended output to CSV would be the icing on the cake. Am I overthinking it? Should it just be a separate script? If so, how?
回答1:
After posting the question with a better description on another thread someone responded with a correct answer. He said that basically what I was seeing is awk buffering output when it's going to a pipeline (since that's lower-overhead), but writing it immediately when it's going to a TTY. He went on to offer a solution by calling fflush() from the awk program.
"Call fflush(), after your print command, add an extra command fflush()."
That fixed it. Thank you all for your efforts.
来源：https://stackoverflow.com/questions/56518422/bash-script-with-awk-output-to-csv