Bash one-liner code to output unique values - bash

Question

I have this command which will output 0, 1 or 2.
This line of code is part of a config file (zabbix), only reason for one-liner code.
mysql -u root -e "show slave status\G" | \
grep -E 'Slave_IO_Running:|Slave_SQL_Running:' | cut -f2 -d':' | \
sed "s/No/0/;s/Yes/1/" | awk '{ SUM += $1} END { print SUM }'
But I want it to output values to be like this so I can setup alert with correct status:
If only Slave_IO_Running is No then output 1.
If only Slave_SQL_Running is No then output 2.
If both are Yes then output 3.
If both are No then output 0.
If no lines/output from show slave status command then output 4.
So something like modify first entry of No with a unique value using sed or awk. And second entry with unique value and so on.
Output of show slave status\G
mysql> show slave status\G
*************************** 1. row ***************************
Slave_IO_State: Waiting for master to send event
Master_Host: 10.10.10.10
Master_User: replicationslave
Master_Port: 3306
Connect_Retry: 60
Master_Log_File: mysql-bin.009081
Read_Master_Log_Pos: 856648307
Relay_Log_File: mysqld-relay-bin.002513
Relay_Log_Pos: 1431694
Relay_Master_Log_File: mysql-bin.009081
Slave_IO_Running: Yes
Slave_SQL_Running: Yes

Answer 1

You can do all the string processing here in awk:
mysql -u root -e "show slave status\G" | awk 'BEGIN {output=0} /Slave_IO_Running.*No/ {output+=1} /Slave_SQL_Running.*No/ {output +=2} END {if(output==3}{print 0} else {if(output==0} {print 3} else {print output}}'
This will start the output counter at 0, if we match Slave_IO_Running with No we'll add 1. If we match Slave_SQL_Running with No we'll add 2, then at the end we'll print the total, which will be 0 if neither are matched, 1 if only IO is No, 2 if only SQL is No and 3 if both are No. Since you want to print 0 if both are Yes we reverse our count at the end, if we got a 3 then both were "No" so print 0, otherwise if it was 0 print 3, else print its own value.

Answer 2

The following awk code could be compacted into a single line if you feel the urge to do that:
awk -F: -v ret=0 '
/Slave_IO_Running:.*No/ { ret=1 }
/Slave_IO_Running:.*Yes/ { yes++ }
/Slave_SQL_Running:.*No/ { ret=(ret==1) ? 0 : 2 }
/Slave_SQL_Running:.*Yes/ { yes++ }
END { print (yes==2) ? 3 : ret }
'
No grep or cut or sed is required, this takes the output of your mysql command directly. It also assumes that Slave_IO_Running will always appear before Slave_SQL_Running in the output of your command.
The notation in the third line and last line functions as an in-line "if" statement -- if the value of ret equals 1, set ret to 0; otherwise set ret to 2.

Answer 3

Whenever you have name to value pairs in your data it's usually clearest, simplest and easiest to enhance later to first create an array mapping the names to the values and then access the values by their names, e.g.:
awk '
{ f[$1]=$2 }
END {
if (f["Slave_10_Running:"] == "Yes")
rslt = (f["Slave_SQL_Running:"] == "Yes" ? 3 : 2)
else
rslt = (f["Slave_SQL_Running:"] == "Yes" ? 1 : 0)
print rslt
}
' file
1