Pick off 'first' dates using cal function in bash

Question

I'm trying to pick off the first Monday, second Tuesday, second Saturday, etc. out of the cal function and piping awk commands. For this month, I can get the first Tuesday correctly with the command below the calendar:

    December 2018
Su Mo Tu We Th Fr Sa
                   1
 2  3  4  5  6  7  8
 9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31

>cal | awk 'NF == 7 && NR > 2 {print $3; exit}'
>4

But this doesn't always work, see the test for March 2016 below:

     March 2016
Su Mo Tu We Th Fr Sa
       1  2  3  4  5
 6  7  8  9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31

>cal 3 2016 | awk 'NF == 7 && NR > 2 {print $3; exit}'
>8

Why does it work for the top example but not the bottom? I would think the bottom example should look at the third line and find the correct result of 1. Is there a fool-proof way to snag these dates using awk -- maybe checking if the column is empty and moving to the next line?

Answer 1

Here's another solution for you

cal | awk 'NR >= 3 { tu = substr($0,7,2) +0 } tu { print tu; exit }'
4

cal 3 2016 | awk 'NR >= 3 { tu = substr($0,7,2) +0 } tu { print tu; exit }'
1

It picks out the Tuesday column (characters 7 and 8), and delivers the first non-zero value. The approach can be safely extrapolated to any day of the week.

Answer 2

With GNU date:

For the first Tuesday of the current month:

for day in 0{1..7}; do [[ "$(date -d $(date +%Y%m${day}) +%a)" = "Tue" ]] && echo "$day";done

For the first Tuesday of a given year and month:

year=2016
month=03
for day in 0{1..7}; do [[ "$(date -d ${year}${month}${day} +%a)" = "Tue" ]] && echo "$day";done

A "Tuesday" variation in awk:

cal | awk -F "" '!/[[:alpha:]]/ && $7$8 ~ /[0-9]/ { print $7$8;exit }'

This checks the hard-coded position of columns 7 and 8 (Tuesday's) in the output (notice the field separator is null, splitting the input on every character) -- after skipping any lines that have letters in them (the header lines). Adjust the 7 & 8 to get other days.

To make it clearer what the previous command is doing, the awk command's fields are numbered per column:

123456789            <--- awk fields $1, $2, $3, ... $9
    December 2018
Su Mo Tu We Th Fr Sa
                   1
 2  3  4  5  6  7  8
 9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31