Animal.txt
January 12, 2021
It’s easiest to approach a task like this in pieces, using a pipeline to gently massage the data to its final form:
awk '
/once was/ { animal = $5 }
/he ate/ { print $4, animal } ' animal.txt |
sort -u |
awk '
NR == 1 { prev = $1; printf "%s %s", $1, $2 }
NR > 1 { if (prev == $1) { printf " %s", $2 }
else { prev = $1; printf "\n%s %s", $1, $2 } }
END { printf "\n" } ' |
awk '
{ if (NR > 1) { print "" }
print "Food:", $1, "Animals who ate it:", NF-1
print "========"
for (N=2; N<=NF; N++) { print $N } } '
The first awk command extracts food/animal pairs from the input. After that command, the pipeline consists of seven lines:
Apples Dog Apples Dog Apples Dog Carrots Dog Carrots Bear Carrots Bear Chicken Bear
The sort command with the -u flag sorts the lines and removes duplicates. Now there are four lines remaining in the pipeline; note that Carrots Bear precedes Carrots Dog, which is the correct order for output:
Apples Dog Carrots Bear Carrots Dog Chicken Bear
The second awk command groups like foods on a single line, using the principle that “output is a weak form of concatenation.” Now three lines remain in the pipeline:
Apples Dog Carrots Bear Dog Chicken Bear
The third awk command produces the final output. Note the attention to the blank line between output groups, which appears before all groups except the first. Here is the final output:
Food: Apples Animals who ate it: 1 ======== Dog Food: Carrots Animals who ate it: 2 ======== Bear Dog Food: Chicken Animals who ate it: 1 ======== Bear
Yes, this is wildly inefficient in terms of computer usage; it spawns four processes and passes lots of data through a pipeline, where it would clearly be possible to do the whole thing in a single awk program. But each piece is tiny, does exactly one thing, is easy to write, and is a candidate for reuse; the grouping program in the second invocation of awk would clearly be useful in other programs, and the first awk program might be useful if other programs read the same input.
You can run the program at https://ideone.com/enCtmf.
A cute little exercise. I hope it’s not counted as cheating if I tackle it using Julia (1.5.2): https://pastebin.com/UDbyNBhQ
I really had no idea that bears ate carrots!
Here’s a solution in Python.
import sys from collections import defaultdict lookup = defaultdict(set) # maps foods to a set of animals with open(sys.argv[1]) as f: for line in f.readlines(): line = line.strip() if line.startswith('The'): # 'The' identifies 'There once ...' animal = line[17:].capitalize() # len('There once was a ') == 17 elif line: food = line[line.index(' ') + 8:].capitalize() # len(' he ate ') == 8 lookup[food].add(animal) for idx, (food, animals) in enumerate(lookup.items()): print(f'Food: {food} Animals who ate it: {len(animals)}') print('=======') for animal in animals: print(animal) if idx + 1 < len(lookup): print()Example usage:
Python
import re
inFile = open(‘animal.txt’)
dt = dict()
l = set()
f = ”
for line in inFile:
if re.match(“There once was a”, line):
if f:
dt[f] = l
f = line.split()[-1].capitalize()
l = set()
elif re.search(“he ate”, line):
s = line.split()[-1].capitalize()
l.add(s)
dt[f] = l
dtr = dict() #reverse dict
for k, v in dt.items():
for iv in v:
if iv in dtr:
dtr[iv].add(k)
else:
dtr[iv] = set([k])
for val in dtr:
print(‘\nFood: {} Animals who ate it: {}’.format(val, len(dtr[val])))
print(‘=======’)
for i in dtr[val]:
print(i)
ouch (
https://pastebin.com/yh2iQweK
Here’s an AWK version. The food names are capitalized, and the output sorted, to match the example.
/There once was a/ { animal = $5 } /he ate/ { foods[$4][animal] = 1 } END { PROCINFO["sorted_in"] = "@val_str_asc" for (f in foods) { printf "Food: %s Animals who ate it: %d\n", raiseCase(f), length(foods[f]) printf "========\n" for (a in foods[f]) { printf "%s\n", a } printf "\n" } } function raiseCase(s) { return toupper(substr(s, 1, 1)) substr(s, 2) }