Script regarding non numerical or empty value

Shell Programming and Scripting
1

Hello Team,

I need bash script to check if my output is non numerical or empty. if its then my output should display default value as 0

basically, I am reading value from txt file. most of numerical value, in case there is no numerical value or its empty, then my output should be 0 .

example like below:

output=`sed '1!d' $SOURCE_FOLDER/$datafile |awk '{print $6}'`
          data="lines1=$output "

          output=`sed '1!d' $SOURCE_FOLDER/$datafile |awk '{print $7}'`
          data+="lines2=$output "

          echo "file value|$date"

Br,Pradeep

2

In lieu of two sed invocations followed by an awk each, you could try a single awk:

awk '{print "lines1=" ($6==$6+0?$6:0) " lines2=" ($7==$7+0?$7:0); exit}' file3
lines1=41531.0 lines2=52529.0
awk '{print "lines1=" ($6==$6+0?$6:0) " lines2=" ($7==$7+0?$7:0); exit}' logfile
lines1=0 lines2=0
1 Like
3

Another approach using awk:

awk '{printf("lines1=%d lines2=%d\n", $6+0, $7+0); exit }' infile

Originally I was going to use NR==1 to match the first line but RudiC's approach using exit should be much more efficient when the input file has a large number of lines.

Replace %d format specifiers with %f , %.2f , etc if you have non-integer values.

1 Like
4

Nice approach! In fact, with the %d format specifier, you don't need the +0 as both an empty or a string field would result in a 0 to be printed:

cf file[1-3]
file1:
/ home k abc txt ABC DEF 
file2:
/ home k abc txt 92 1023
file3:
AB CD EF
awk '{printf("lines1=%d lines2=%d\n", $6, $7); exit }' file1
lines1=0 lines2=0
awk '{printf("lines1=%d lines2=%d\n", $6, $7); exit }' file2
lines1=92 lines2=1023
awk '{printf("lines1=%d lines2=%d\n", $6, $7); exit }' file3
lines1=0 lines2=0

The only drawback is when a field has a string starting with numeric characters like

cat file4
A B C D E F23 24G H

Your approach:

awk '{printf("lines1=%d lines2=%d\n", $6+0, $7+0); exit }' file4
lines1=0 lines2=24

My approach:

awk '{print "lines1=" ($6==$6+0?$6:0) " lines2=" ($7==$7+0?$7:0); exit}' file4
lines1=0 lines2=0

The OP needs to decide which one would satisfy his/her needs.

5

Thanks all for reply, I have quick question here. how about having multiple lines in single file like below.
.

e.g
100 4  5 101 400
 900 1245 789 P

I want to read multiple lines from files and display output related each field.
my output should be like this.

process1= 100
process2 = 900
process3 = 789 
process4 = 0

process4 will be 0 since its having non numerical value.
6

Awk has a built in variable NR which matches the row number of the file so you can test this when extracting your process numbers:

awk '
NR==1 { printf("process1= %d\n", $1+0) }
NR==2 { 
   printf("process2 = %d\n", $1+0)
   printf("process3 = %d\n", $3+0)
   printf("process4 = %d\n", $4+0)
   exit
}' infile