Holiday table make

Question

I need script to convert this holiday list into date format. My input file

MMYYYY CAL_B2K_ID   123456789012345678901234567890  
------ ------------ ------------------------------- 
012016 821          YY      Y      Y      Y      Y  
012016 DC           YY      Y      Y      Y      Y  
022016 DC                Y      Y      Y      Y     
022016 821               Y      Y      Y      Y     
032016 DC               Y      Y      Y      Y      
032016 821              Y      Y      Y      Y      
042016 821           Y      Y      Y      Y         

Now 123456789012345678901234567890 is like date when first time 0 comes it is 10 and then 1 means 11 and this is so on. ...

 1 2 3 4 5 6 7 8 9 10 11 12 13 14............

So I have created script where ever Y is present it convert into date suppose in first line Y is present under 1, 2,9,6,3,9

In my script i have cut each line from column 21- 52 and checking if that column is Y, if yes then it covert into date like if Y is present under column number 23. Then I am subratcing (coloumn number - 20) it will give me 3 And then i am using concat operatior to make date.

Below is my script. But i am facing issue when there is no Y under any coloumn it takes my var as blank and blank is also coming under true condition .

var=echo $f1 | cut -c$i
if [[ "$var"='Y' ]] ;

And my script is below.

 rm f1.txt
 set -x
 while read f1
 do
 for((i=21;i<23;i++));
 do
 var=`echo $f1 | cut -c$i`
 if [[ "$var"='Y' ]] ;
 then
 echo $var
 month=`echo $f1 | cut -c1-2`
 year=`echo $f1 |cut -c3-6`
 date=$(($i-20))
 echo $year"-"$month"-"$date >> f1.txt
 fi
done
done < holiday_india.txt`

Please let me know what I will do so that when there is nothing in var , it will not come under If condition true.

O/p of this script is that it convert all coloumn of each line from 21-52 into date not where Y is present.

Hope I clarify now.

Answer 1

Okay, I copied your script below, and fixed the indentation so that it's a bit easier for me to read. (And I also changed the backticks to $() and removed the fixed file names because I like it that way.)

while read f1 ; do
    for((i=21;i<23;i++)); do
        var=$(echo $f1 | cut -c$i)
        if [[ "$var"='Y' ]] ; then
            echo $var
            month=$(echo $f1 | cut -c1-2)
            year=$(echo $f1 |cut -c3-6)
            date=$(($i-20))
            echo "$year-$month-$date"
        fi
    done
done

Two things jump up:

if [[ "$var"='Y' ]] ; then

This doesn't work; it needs spaces around the equal sign. Try and see what [[ X=Y ]] && echo foo does.

Also, here

var=$(echo $f1 | cut -c$i)

f1 contains something like 012016 821 YY Y Y Y Y, but since it’s not quoted, it gets expanded, then word-split along the spaces, after which echo prints all the words separated by single spaces, resulting in 012016 821 YY Y Y Y Y. You could fix it by adding the quotes: echo "$f1" | cut -c$i, but there's also "${f1:N:L}" to grab a substring of length L from position N. It is zero-indexed, though, so the range of $i will need to be modified.

Then,

echo "$year-$month-$date"

This could be changed to printf "%04d-%02d-%02d\n" "$year" "$month" "$date" to make sure the date is always printed with two digits, and to protect against possible anomalous values of $year.

Might as well move the month and year assignments to the outer loop, since they stay the same for the whole line.

Now we have:

while read f1; do
    month="${f1:0:2}"
    year="${f1:2:4}"
    for (( i=20 ; i < 20 + 31 ; i++ )); do
        var="${f1:$i:1}"
        if [[ "$var" = "Y" ]] ; then
            date=$(( $i - 19 ))
            printf "%04d-%02d-%02d\n" "$year" "$month" "$date"
        fi
    done
done

The output looks about correct when I run it with bash holiday.sh < holiday_india.txt:

2016-01-01
2016-01-02
2016-01-09
2016-01-16
2016-01-23
2016-01-30
2016-01-01
    ︙

Of course you seem to have all months doubled, so they are doubled in the output too. Perhaps add a test for the contents of the second column, or run it with something like this to check only lines with the given id:

grep DC holiday_india.txt | bash holiday.sh

(This would of course also skip the header lines. Or you could check that the month/year is sane at the beginning of the line.)

_{This was tagged with ksh, but I used bash. The copy of ksh I have supports ${foo:N:L}, which I think was the only new syntax I used.}