def intish(str):
try:
return int(str)
except:
return 0
Can python's default be reproduced as easily in awk?2. You'd insert field = line.split('\t') at the beginning of the loop and then refer to field[2]
3. os.popen or subprocess.run
I buy the "less cluttered" argument when the problem matches awk's defaults. I vehemently disagree with the "more flexible" argument. A problem perfectly suited to awk can easily turn to a poor fit with the addition of a single, seemingly innocuous requirement (e.g. in your subprocess example, log the standard error of your subprocess into a separate file).
!/^[0-9]+$/ {
print "invalid input: " $0 > "/dev/stderr"
exit 1
}
at the beginning of the script. None of the other actions need to be changed; but with your implementation, all of the calls to "int" need to be changed to "intish".I've got the following script (I stopped playing games with line breaks):
#!/usr/bin/env gawk -f
BEGIN {
FS = "|"
}
$2 % 3 == 0 {
printf("Fizz")
replaced = 1
}
$2 % 5 == 0 {
printf("Buzz")
replaced = 1
}
replaced {
replaced = 0
printf("\n")
next
}
{
system("cal " $2 " 2018 2> errors.txt")
}
Which can produce the following output: $ ./script.awk <<EOF
> thing1|0
> thing2|3
> thing3|7
> thing4|13
> EOF
FizzBuzz
Fizz
July 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
$ cat errors.txt
cal: 13 is neither a month number (1..12) nor a name
- What does the equivalent program in Python look like?- How many characters does it have with respect to the number of characters in the awk script? (259 with shebang).
- How many characters would need to change to split by "," instead? (1 for awk). (You can achieve this in Python, but you'll end up spending characters on a utility function.)
- How many characters would need to be added to print "INVALID: " and then the input value for lines with non-numeric values in the second column, then skip to the next line? (55 for awk)
Character adds/changes are the best proxy for "flexibility" I could think of that doesn't go far afield into static code analysis.
I love Python and don't think awk is a good solution for extremely large or complex programs; however, it seems obvious to me that it is significantly more flexible than Python in every line-oriented text-processing task. The combination of opinionated assumptions, built-in functions and automatically-set variables, and the pattern-action approach to code organization, all add up to a powerful tool that's still worth using in order to keep tasks from becoming large or complex in the first place.
2. How would you modify it so it parsed a tab-delimited file and did FizzBuzz on the third column? With awk it is a simple matter of setting FS="\t" and changing $0 to $3?
3. How would you modify it so instead of being output unmodified, rows with $3 that are neither fizz nor buzz output the result of a subprocess called with the second column's contents?
Now you might say that this is all goalpost-moving, but that's the point. AWK is more flexible and less cluttered in situations where the goalposts tend to get moved, but where the basic text processing paradigm stays the same.