i have a data.csv file that i want to convert to txt and then find how many even numbers are contained:

my data.csv is like this:

# Recorded data from sensor XaB.v2
# Recording started on Oct 29 15:00:00
# Recording ended on Oct 29 15:15:59
# X, Y, Z, X', Y', Z'
55, 14, 48, 62, 78, 41
32, 52, 94, 11, 17, 83

and so on.

i tried this way but maybe there is a better way

grep -v '^#' data.csv | sed -e 's/,//g' > data.txt

grep -o '[0-9]\+' data.txt | awk '$1%2 == 0' | wc -l

expecially the sed command i don't know if this is the optimal way.

Then i have to create n copies of my data.txt with scaled values divided by i, again i tried with this script but i do not know if there is a simpler and cleaner way to do it:

n=$1

for ((i=1; i<=n; i++)) do
    awk -v divisor="$i" '{ for (j=1; j<=NF; j++) printf "%.6f", $j / divisor; print "" }' data.txt > "data_${i}.txt"
done

I have not included the necessary checks (syntax and n as a positive integer) in this snippet.

i have a data.csv file that i want to convert to txt and then find how many even numbers are contained:

my data.csv is like this:

# Recorded data from sensor XaB.v2
# Recording started on Oct 29 15:00:00
# Recording ended on Oct 29 15:15:59
# X, Y, Z, X', Y', Z'
55, 14, 48, 62, 78, 41
32, 52, 94, 11, 17, 83

and so on.

i tried this way but maybe there is a better way

grep -v '^#' data.csv | sed -e 's/,//g' > data.txt

grep -o '[0-9]\+' data.txt | awk '$1%2 == 0' | wc -l

expecially the sed command i don't know if this is the optimal way.

Then i have to create n copies of my data.txt with scaled values divided by i, again i tried with this script but i do not know if there is a simpler and cleaner way to do it:

n=$1

for ((i=1; i<=n; i++)) do
    awk -v divisor="$i" '{ for (j=1; j<=NF; j++) printf "%.6f", $j / divisor; print "" }' data.txt > "data_${i}.txt"
done

I have not included the necessary checks (syntax and n as a positive integer) in this snippet.

• bash
• awk

2 Answers 2

You can use awk to count the even numbers. Use a condition in the awk script to filter out comment lines, rather than piping from sed.

awk -F', ' '!/^#/ {for (i = 1; i <= NF; i++) if ($i % 2 == 0) counter++}
            END {print counter}' < data.csv

Recent versions of GNU awk can parse the CSV format (-k or --csv option). For the first goal (count even values):

awk -k '
!/^#/ {
  for(i = 1; i <= NF; i++) s += 1 - ($i % 2)
}
END {
  print s
}' data.csv

For the second goal we can also let awk run the outermost loop:

awk -v n="$1" -k '
!/^#/ {
  for(i = 1; i <= n && f = "data_" i ".txt"; i++)
    for(j = 1; j <= NF; j++)
      printf("%s%.6f%s", j==1 ? "" : " ", $j/i, j==NF ? "\n" : "") > f
}' data.csv

(spaces added to separate the output values). Note that with simple CSV like this (no multi-line fields, no quoted fields with commas...), even with older or non GNU versions of awk, as shown in Barmar's answer, we could use -F ',' and obtain the same result as with -k.