script functionality

[Dave Walley]

Hi.

I have a problem whereby I require to extract values from a file. The problem is some lines contain multiple values and I need to retrieve just the first or second value on occasions.

e.g.
Buffer Nowait %: 100.00 Redo NoWait %: 70.00
Buffer Hit %: 96.68 In-memory Sort %: 99.83
Library Hit %: 99.89 Soft Parse %: 99.64
Execute to Parse %: 57.07 Latch Hit %: 99.99

From this I need to retrieve each value in turn, I am using grep to isolate the row but now I need to extract only the relevant value, e.g. when I need the IN_MEMORY SORT I only want 99.83 and not the entire line. Similarly when I want BUFFER HIT I want 99.68 and nothing else from the line.

Any ideas?

Thanks in advance

Dave

[Tore_1]

all the numbers :

awk -F': ' '{print $2,$3}' file | awk '{print $1,$NF}'

[Colin Topliss]

To get the first value:

cat myfile |awk -F% '{print $2}' |awk '{print $2}'

To get the second value:

cat myfile |awk -F% '{print $2}' |awk '{print $2}'

[Steve Steel]

hi

cat file|while read line
do
treatment of line
done


Then you have the lines 1 by 1 in varaible line.

ex

cat file|while read line
do
echo $line|cut -f3 -d":"
done

gives only the last field.


What do you need for format


Steve Steel

[H.Merijn Brand (procura]

a5:/tmp 110 > cat xx Buffer Nowait %: 100.00 Redo NoWait %: 70.00
Buffer Hit %: 96.68 In-memory Sort %: 99.83
Library Hit %: 99.89 Soft Parse %: 99.64
Execute to Parse %: 57.07 Latch Hit %: 99.99
a5:/tmp 111 > perl -nle'BEGIN{$pat=shift}/\s*([a-z_. -]*?$pat[\w%:. -]*?[\d.]+)/i&&print$1' memory xx
In-memory Sort %: 99.83
a5:/tmp 112 > perl -nle'BEGIN{$pat=shift}/\s*([a-z_. -]*?$pat[\w%:. -]*?[\d.]+)/i&&print$1' parse xx
Soft Parse %: 99.64
Execute to Parse %: 57.07
a5:/tmp 113 >

you can allways turn this into a script :)

Enjoy, have FUN! H.Merijn

[Steve Steel]

Hi

awk is fun but anyway here is another silly one


cat file|while read line
do
f1=$(echo $line|cut -f1 -d":")
f2=$(echo $line|cut -f2 -d":")
f2a=$(echo $f2|cut -f1 -d" ")
f2b=$(echo $f2|cut -f2 -d" ")
f3=$(echo $line|cut -f3 -d":")
echo $f2a $f1
echo $f3 $f2b
done|sort -n


output is

57.07 Execute to Parse %
70.00 Redo
96.68 Buffer Hit %
99.64 Soft
99.83 In-memory
99.89 Library Hit %
99.99 Latch
100.00 Buffer Nowait %

It is endless

Steve steel

[H.Merijn Brand (procura]

Since Opera ruined my newlines anyway, here's the script approach:

a5:/tmp 117 > cat xx
Buffer Nowait %: 100.00 Redo NoWait %: 70.00
Buffer Hit %: 96.68 In-memory Sort %: 99.83
Library Hit %: 99.89 Soft Parse %: 99.64
Execute to Parse %: 57.07 Latch Hit %: 99.99
a5:/tmp 118 > cat xx.pl
#!/usr/bin/perl -l
$pat = shift;
while (<>) {
/\s*([a-z_. -]*?$pat[\w%:. -]*?[\d.]+)/i && print $1;
}
a5:/tmp 119 > xx.pl memory xx
In-memory Sort %: 99.83
a5:/tmp 120 > xx.pl parse xx
Soft Parse %: 99.64
Execute to Parse %: 57.07
a5:/tmp 121 > xx.pl soft xx
Soft Parse %: 99.64
a5:/tmp 122 >

Enjoy, have FUN! H.Merijn

[Robin Wakefield]

Hi Dave,

Here's another awk solution:

Create an awk script:

=======================
$0 ~ var{
p=index($0,var)
for (i=p+length(var)+4;i{printf substr($0,i,1)}
}END{print}
=======================

then run using:

# awk -f awkscript var="Soft Parse" yourfile
99.64

# awk -f awkscript var="Buffer Nowait" yourfile
100.00

rgds, Robin