HackerRank Shell

Intro

Unless otherwise noted, assume all scripts contain the following shebang:

#!/usr/bin/env bash

Easy Challenges

https://www.hackerrank.com/domains/shell

Let’s Echo

Tags: #cmdline #shell #bash #echo #printf Links: challenge

$ echo HELLO

$ printf '%s\n' HELLO

Looping With Numbers

  • Tags: #cmdline #shell #bash #numbers #looping #for

  • Links: challenge

for (( i = 1; i <= 9; ++i ))
do
  echo "$i"
done

Or using ranges:

$ printf '%d\n' {1..50}

Looping And Skipping

  • Tags: #cmdline, #numbers #looping #for

  • Links: challenge

for (( i = 1; i <= 9; ++i ))
do
  if (( i % 2 == 0 ))
  then
    continue
  fi
  echo "$i"
done
$ bash script.sh
1
3
5
7
9

Could also use ‘’echo:’’

$ echo -ne {1..9..2} '\n'

The -e option is to enable some escapes. help echo for more.

Or using seq:

$ seq -s ' ' 1 2 9

A Personalized Echo

  • Tags: #cmdline #read #echo

  • Links: challenge

$ read -r name
$ printf 'Welcome %s\n' "$name"

The World of Numbers

  • Tags: #cmdline #shell #bash #numbers #math #bc #ranges

  • Links: challenge

First, see this clever use of range to produce the math expressions:

$ read -r x y
8 2

$ printf '%s\n' "$x"{+,-,*,/}"$y"
8+2
8-2
8*2
8/2

Then, feed those expressions to bc:

$ read -r x y
8 2

$ printf '%s\n' "scale=2; $x"{+,-,*,/}"$y" | bc
10
6
16
4.00

If y is negative, like -2 we would receive an error:

$ read -r x y
5 -2

$ printf '%s\n' "scale=2; $x"{+,-,*,/}"$y" | bc
3
(standard_in) 2: syntax error
-10
-2.50

Adding parenthesis prevents the error, because our expression would be like 5—​2, but 5-(-2) is OK with bc:

$ read -r x y
5 -2

$ printf '%s\n' "scale=2; $x"{+,-,*,/}"($y)" | bc
3
7
-10
-2.50

Or something more manual and verbose:

read x </dev/stdin
read y </dev/stdin

printf '%d\n' $(( x + y ))
printf '%d\n' $(( x - y ))
printf '%d\n' $(( x * y ))
printf '%d\n' $(( x / y ))
The challenge wants integer division, so, we simply omit bc’s scale special variable. 
read -r answer

case "$answer" in
  [Yy]*)
    printf '%s\n' YES
    ;;
  [Nn]*)
    printf '%s\n' NO
    ;;
  *)
    printf '%s\n' 'What the poop‽ 💩'
    ;;
esac
$ bash script.sh
yes
YES

$ bash script.sh
Y
YES

$ bash script.sh
n
NO

$ bash script.sh
lol
What the poop‽ 💩

Getting started with conditionals

  • Tags: #cmdline #shell #bash #conditionals

  • Links: challenge

read -r answer

case "$answer" in
  [Yy]*)
    printf '%s\n' YES
    ;;
  [Nn]*)
    printf '%s\n' NO
    ;;
  *)
    printf '%s\n' 'What the poop‽ 💩'
    ;;
esac
$ bash script.sh
yes
YES

$ bash script.sh
Y
YES

$ bash script.sh
n
NO

$ bash script.sh
lol
What the poop‽ 💩

More on Conditionals

  • Tags: #cmdline #shell #bash #conditionals #math

  • Links: challenge

Solution based on side lengths.

  • equilateral: x == y && y == z

  • scalene: x != y && y != z && z != x

  • isosceles: any other

read -r x
read -r y
read -r z

[[ "$x" == "$y" ]] && [[ "$y" == "$z" ]] && echo EQUILATERAL && exit 0
[[ "$x" != "$y" ]] && [[ "$y" != "$z" ]] && [[ "$z" != "$x" ]] && echo SCALENE && exit 0
echo ISOSCELES && exit 0

Arithmetic Operations

  • Tags: #cmdline #shell #bash #math #bc

  • Links: challenge

expression="$1"
printf '%.3f\n' "$(echo "$expression" | bc -l)"

bc -l produces up to 6 decimal places. If we use bc scale to 3, for instance, depending on the result, we would produce wrong results because printf %f format specifier does rounding by itself.

bc scale is 0 by default if not explicitly set. Also, bc does no rounding.

printf rounds up from 6, and down from 5:

$ printf '%.3f\n' 1.2583
1.258
$ printf '%.3f\n' 1.2585
1.258
$ printf '%.3f\n' 1.2586
1.259

Only when the number after 8 passes 5, that is, 6 and above, is that the number is rounded up to 1.259. If one uses scale=3 in bc, then it truncates (does not round) to three decimal places and printf has no way to round up, making the solution to the exercise incorrect. Therefore, we use bc -l without scale, or use scale=4 at least.

Compute the Average

  • Tags: #cmdline #shell #bash #math

  • Links: challenge

read -r n
sum=0

if [[ "$n" == 0 ]]
then
  printf '%.3f\n' "$(echo 'scale=4; 0' | bc -l)"
  exit 0
fi

for ((i = 0; i < n; ++i))
do
  read -r x
  sum=$((sum + x))
done

printf '%.3f\n' "$(echo "scale=4; $sum / $n" | bc -l)"

We used scale=4 by the same reasons described earlier about truncating and rounding.

cut Challenges

  • Tags: #cmdline #shell #bash #cut

$ cut -b 3 -

$ cut -b 2,7 -

$ cut -b 2-7 -

$ cut -b 1-4 -

$ cut -d $'\t' -f 1,2,3 -

$ cut -c 13- -

$ cut -d ' ' -f 4 -

$ cut -d ' ' -f 1,2,3 -

$ cut -d $'\t' -f 2- -

Head of Text File Challenges

$ head -n 20

$ head -c 20

Middle of a Text File

  • Tags: #cmdline #shell #bash #sed

  • Links: challenge

$ sed -n '12,22 p'

Tail of a Text File 1 and 2

  • Tags: #cmdline #shell #bash #tail

  • Links: challenge

$ tail -n 20 -

$ tail -c 20 -

tr Command 1

  • Tags: #cmdline #shell #bash #tr #here-document #assignment

  • Links: challenge

# Assign some text to the variable `input'.
$ read -r -d '' input << 'EOF'
int i = (int) 5.8;
int res = (23 + i) * 2;
EOF

# Inspect `input' contents.
$ echo "$input"
int i = (int) 5.8;
int res = (23 + i) * 2;

# Apply `tr' to `input' and see ( and ) replaced with [ and ].
$ echo "$input" | tr '()' '[]'
int i = [int] 5.8;
int res = [23 + i] * 2;

A Here Document is used to assign lines of text to the variable input.

tr Command 2

  • Tags: #cmdline #shell #bash #tr

  • Links: challenge

$ tr -d 'a-z'

tr Command 3

  • Tags: #cmdline #shell #bash #tr

  • Links: challenge

$ tr -s ' '

sort Lines Challenges

  • Tags: #cmdline #shell #bash #sort

  • Links: challenge

$ echo -e 'aa\nbb\naa\ncc\nff\ncc' | sort -
aa
aa
bb
cc
cc
ff

$ echo -e 'aa\nbb\naa\ncc\nff\ncc' | sort -r -
ff
cc
cc
bb
aa
aa

$ echo -e '2.1\n3\n0.2\n0' | sort -n -
0
0.2
2.1
3

$ echo -e '2.1\n3\n0.2\n0' | sort -nr -
3
2.1
0.2
0

# Sort by field 2, taking Tab as field separator.
$ sort -t $'\t' -nr -k 2 -

# Same, but in ascending order.
$ sort -t $'\t' -n -k 2 -

# This time the delimiter is a “|” character
$ sort -t '|' -nr -k 2 -

uniq Challenges

  • Tags: #cmdline #shell #bash #uniq

  • Links: challenge

$ uniq -
​```

Display the count of lines that were uniqfied and the uniqfied lines without leading whitespace/tabs:

$ read -r -d '' lines << 'EOF'
> foo
> foo
> bar
> bar
> bar
> tux
> EOF

$ echo "$lines" | uniq -c - | sed 's/ \+\([0-9]\+ [^ ]\+\)/\1/'
2 foo
3 bar
1 tux

$ echo "$lines" | uniq -c - | sed 's/^[[:space:]]*//g'
2 foo
3 bar
1 tux

$ echo "$lines" | uniq -c - | cut -b 7- -
2 foo
3 bar
1 tux

$ echo "$lines" | uniq -c - | xargs -l
2 foo
3 bar
1 tux

$ echo "$lines" | uniq -c - | xargs -L 1
2 foo
3 bar
1 tux

$ echo "$lines" | uniq -c - | colrm 1 6
2 foo
3 bar
1 tux

# Case Insenstivie.
$ read -r -d '' lines << 'EOF'
> FoO
> fOO
> baR
> Bar
> bAr
> TUX
> EOF

$ echo "$lines" | uniq -ci - | cut -b 7- -
2 FoO
3 baR
1 TUX

$ echo "$lines" | uniq -u -
TUX

Read In An Array

  • Tags: #cmdline #shell #bash #arrays

  • Links: challenge

$ arr=()
$ while read -r line ; do arr+=("$line") ; done < /dev/stdin
$ echo "${#arr[*]}"

Display an Element of an Array

  • Tags: #cmdline #shell #bash #arrays

  • Links: challenge

mapfile -t countries
echo "${countries[3]}"

-t in mapfile removes the trailing delimiter so the array elements are “clean”.

Count Elements in an Array

  • Tags: #cmdline #shell #bash #arrays

  • Links: challenge

mapfile -t countries
echo "${#countries[@]}"

Slice An Array

  • Tags: #cmdline #shell #bash #arrays

  • Links: challenge

Print the array with the syntax ${arr[*]:OFFSET:LENGTH}.

$ read -r -d '' countries << 'EOF'
> Namibia
> Nauru
> Nepal
> Netherlands
> NewZealand
> Nicaragua
> Niger
> Nigeria
> NorthKorea
> Norway
> EOF

$ echo "${arr[*]:3:5}"
Netherlands NewZealand Nicaragua Niger Nigeria NorthKorea Norway

Could read with countries=($(cat)) too, but ShellSheck complains. Either use the read as above, or with mapfile -t arr.

Other options would be:

paste -d ' ' -s | cut -d ' ' -f4-8 -

and:

head -8 | tail -5 | paste -s -d ' ' -

Concatenate Array With Itself

  • Tags: #cmdline #shell #bash #arrays

  • Links: challenge

mapfile -t countries

countries+=("${countries[@]}" "${countries[@]}")

echo "${countries[*]}"

grep A

  • Tags: #cmdline #shell #sed

  • Links: challenge

$ grep -iw 'th\(e\|at\|en\|ose\)'

grep B

  • Tags: #cmdline #shell #grep

  • Links: challenge

Works locally but not on HackerRank:

$ grep '\(.\) \?\1'

This works locally and on HackerRank:

$ grep '\(.\) \?\1'

sed 3

  • Tags: #cmdline #shell #sed

  • Links: challenge

$ sed 's/[Tt][Hh][Yy]/{&}/g'

sed 4

  • Tags: #cmdline #shell #sed

  • Links: challenge

$ sed 's/.* \([0-9]\{4\}\)/**** **** **** \1/g'

Or

$ sed 's/[0-9]\+ /**** /g'

Medium Challenges

Paste 1

  • Tags: #cmdline #shell #paste

  • Links: challenge

$ paste -s -d ';' -

paste 2

paste -d ';' - - -

paste 3

  • Tags: #cmdline #shell #paste

  • Links: challenge

$ paste -s -

paste 4

  • Tags: #cmdline #shell #paste

  • Links: challenge

$ paste - - -

sed 1

  • Tags: #cmdline #shell #sed

  • Links: challenge

$ sed 's/\<the\>/this/'

sed 2

  • Tags: #cmdline #shell #sed

  • Links: challenge

grep challenges

$ grep '\<the\>'

$ grep -i '\<the\>'

$ grep -iv '\<that\>'

awk challenges

Challenge 1:

$ awk '{ if ($4 == "") print "Not all scores are available for " $1 }'

Challenge 2:

awk '{
  answer[0] = "Fail";
  answer[1] = "Pass";
  print $1, ":", answer[$2 >= 50 && $3 >= 50 && $4 >= 50];
}'

Challenge 3:

awk '{
  avg=($2 + $3 + $4) / 3
  if (avg >= 80)
    print $0 " : A";
  else if (avg >= 60)
    print $0 " : B";
  else
    print $0 " : FAIL";
}'

Challenge 4:

awk 'ORS=NR % 2 ? ";" : "\n"'

Filter an Array With Patterns

  • Tags: #cmdline #shell #bash #arrays #pattern-matching

  • Links: challenge

while read -r line ; do
  if [[ ! "$line" =~ [Aa] ]]
  then
    echo "$line"
  fi
done

Remove First Capital Letter From Each Array Element

  • Tags: #cmdline #shell #bash #arrays #pattern-matching

  • Links: challenge

arr=()

while read -r line ; do
  arr+=("${line/[A-Z]/.}")
done

echo "${arr[*]}"

Hard Challenges

sed 5

  • Tags: #cmdline #shell #sed

  • Links: challenge

sed 's/\([0-9]\+\) \([0-9]\+\) \([0-9]\+\) \([0-9]\+\)/\4 \3 \2 \1/'
Backreferences in the search pattern mean they match the same chars, not the same general regex. That is, (.)o(.) matches “bob” or “bob”, for instance, but not “bop”. If (.) matched “x”, then \1 in the search must also match an “x”. That is why we can’t do s/\([0-9]\+\) \1 \1 \1, because it would only match if all four fields of the number were the same thing, like “1234 1234 1234 1234”.

Lonely Integer

  • Tags: #cmdline #shell #bash #numbers

  • Links: challenge

Not very elegant, but makes use of arrays, which is what they ask for.

#!/usr/bin/env bash

#
# This solution uses a histogram-like approach.
#

# Dummy-read, since we don't need the first argument they
# feed into the input.
read -r

# Read input numbers.
read -r -a nums

# An array to keep track of which numbers appeared how many times.
declare -A hist

for n in "${nums[@]}"
do
  if [[ -z "${hist[$n]}" ]]
  then
    # Use the number as index and increment that index and
    # initialize it to 1.
    hist[$n]=1
  else
    # Increment it each time that number appears.
    hist[$n]=$((${hist[$n]} + 1))
  fi
done

# Iterate over the indexes.
for idx in "${!hist[@]}"
do
  # If that number appeared only once...
  if (( hist[$idx] == 1 ))
  then
    # ...then print it and bail out.
    echo "$idx"
    break;
  fi
done

Fractal Tree

  • Tags: #cmdline #shell #bash

  • Links: challenge

#!/usr/bin/env bash

#
# Invoke it like this:
#
#   bash script.sh 5
#

declare -A grid
rows=63
cols=100

#
# Initialize the 63x100 grid with underscores.
#
init () {
  for (( row = 0; row < rows; ++row ))
  do
    for (( col = 0; col < cols; ++col ))
    do
      grid[$row,$col]=_
    done
  done
}

#
# Actually treeify the drawing.
#
treeify () {
  local count=$1
  local row=$2
  local col=$3
  local iteration=$4

  for (( i = 0; i < count; ++i ))
  do
    grid[$row,$col]=1
    (( row -= 1 ))
  done

  for (( i = 0; i < count; i++ ))
  do
    grid[$row,$((col - i - 1))]=1
    grid[$row,$((col + i + 1))]=1
    (( row -= 1 ))
  done

  if (( iteration > 1 ))
  then
    treeify $(( count >> 1 )) "$row" $(( col - count )) $(( iteration - 1 ))
    treeify $(( count >> 1 )) "$row" $(( col + count )) $(( iteration - 1 ))
  fi

}

#
# Simply output the grid, already treeified, to the screen.
#
display () {
  for (( row = 0 ; row < rows ; ++row ))
  do
    for (( col = 0 ; col < cols ; ++col ))
    do
      printf '%s' "${grid[$row,$col]}"
    done
    printf '\n'
  done
}

initial_count=16
initial_row=62
initial_col=49
iterations="${1:-5}"

if (( 1 > iterations || iterations > 5 ))
then
  printf '%s\n' 'Provide a number between 1 and 5, please.' 1>&2
else
  init
  treeify "$initial_count" "$initial_row" "$initial_col" "$iterations"
  display
fi

The End