KSH: Loop performance - performance

Question

I need to process a file with approximately 120k lines that has the following format using ksh:
"[UserId=USER1]";"Client=001";"Locked_Status=0";"TYPE=A";"Last_Logon=00000000";"Valid_To=99991231";"Password_Change=20120131";"Last_Password_Change=29990"
"[UserId=USER2]";"Client=000";"Locked_Status=0";"TYPE=A";"Last_Logon=20141020";"Valid_To=00000000";"Password_Change=20140620";"Last_Password_Change=9501"
"[UserId=USER3]";"Client=002";"Locked_Status=0";"TYPE=A";"Last_Logon=00000000";"Valid_To=99991231";"Password_Change=20140304";"Last_Password_Change=9817"
The output should be something like:
[UserId=USER1] Client=001
Locked_Status=0
TYPE=A
Last_Logon=00000000
Valid_To=99991231
Password_Change=20120131
Last_Password_Change=29985
[UserId=USER2]
Client=000
Locked_Status=0
TYPE=A
Last_Logon=20141020
Valid_To=00000000
Password_Change=20140620
Last_Password_Change=9496
[UserId=User3]
Client=002
Locked_Status=0
TYPE=A
Last_Logon=00000000
Valid_To=99991231
Password_Change=20140304
Last_Password_Change=9812
I initially used the following code do process the file:
for a in $(<$1)
do
a=$(echo $a|sed -e 's/;/ /g' -e 's/"//g')
for b in $a
do
print $b
done
done
It was taking around 3hrs to process 120k lines.
Then I tried to improved the code changing it to the following:
for a in $(<$1)
do
printf "\n$(echo $a|sed -e 's/"//g' -e 's/;/\\n/g')"
done
That gave me 2hrs processing time however it still takes too long to process 120k lines
At last I tried this code which processed the 120k lines in 3secs!
perl -ne '
chomp;
s/\"//g;
s/;/\n/g;
print;
' <$1
Is there anyway I can improve the code in KSH to achieve similar performance? I believe that I must be missing something in my KSH code... Help me to find out please.
Thanks in advance

Answer 1

How about:
tr ';' '\n' < file | tr -d '"'
Your code is assigning every whitespace-delimited word to variable "a" in turn, and thus invoking sed once for each word in the file. Clearly a lot of accumulated overhead spawning all those processes. The idiom to iterate over the lines of a file is:
while IFS= read -r line; do ...; done < file

Answer 2

You suggestion worked perfectly
Host> wc -l /tmp/MyTest
114449 /tmp/MyTest
Host> time tr ';' '\n' < /tmp/MyTest | tr -d '"' > /tmp/zuza.out
real 0m1.04s
user 0m1.06s
sys 0m0.08s
Host> time perl -ne '
chomp;
s/\"//g;
s/;/\n/g;
print "\n$_";
' </tmp/MyTest > /tmp/zuza
real 0m1.30s
user 0m0.60s
sys 0m0.08s