Arrange a Haft-Sin table

Question

It's Nowruz and you want to set up a Haft-Sin table by words. This means finding seven words that start with letter s.

The Challenge

Write a program which its input is a list of words separated by space, and output at most first 7 words which starts by letter s. If the s is before letter h it doesn't count because it would not pronounce /s/.

Input

An arbitrary length string containing words separated by space. Words must not contain anything other than letters (uppercase or lowercase) and numbers and _.

These inputs are valid:

hello Puzzle code_golf 12
Start say_hello separating_by_space_is_right
I am a valid word list

And these inputs are invalid:

code-golf, #invalid_word, separating_by_comma_is_wrong
I'm an invalid word list

Output

The first 7 words which starts by letter S-s and not followed by letter H-h, in every acceptable way (comma separated, space separated, new-line etc) and in any order.

• If two words are duplicate don't count them twice. Every single word is count once.
• If the input contains less that 7 word starting with s output nothing. Don't output the words.
• The output must contain the exact word which is in the input. So if the input contains SuPER, output should be SuPER and not SUPER or super or any other form of lower and upper case.
• Words pronunciation matter. The word Speed and SPEED both count the same. You may want to lowercase all the input and unique the words and then check for words.

test-cases

input:

speed speed new car book seven sad sum power fun super sister silver silly start

output:

speed seven sad sum super sister silver

input:

speed SpEEd new book seven sad sum power fun super sister silver silly start

output:

speed seven sad sum super sister silver

input:

sheep speed new car book seven sad sum power fun super sister silver silly start

output:

speed seven sad sum super sister silver

input:

first second third

output:

Edited

This was my first question and I missed many special cases. I try to clarify them.

Answer 1

Vyxal, 28 bytes

⌈'⇩ḣh\h≠$h\s=∧;:ɽ:vḟUİ₅6>7*Ẏ

Try it Online! or Try some testcases! Mis-deduplication should be fixed now.

⌈'⇩ḣh\h≠$h\s=∧;:ɽ:vḟUİ₅6>7*Ẏ
⌈'             ;                 Filter input, split on spaces, by:
  ⇩ḣ                                 Push (input lowercased)([0], [1:])
    h\h≠                             Second character (input[1:][0]) isn't `h`
        $h\s=∧                       and first character isn `s`

                :ɽ               Duplicate, lowercase each
                  :vḟ            Find first occurrences of each of ^ in ^
                     U           Uniquify ^
                      İ          Index ^ into filtered list, resulting
                                 in the properly deduplicated list

                       ₅6>       1 if len(^) is 7 or more, otherwise 0
                          7*Ẏ    Multiply by 7 and slice [0:that]

Answer 2

R, 57 61 58 79 bytes

Edit: -2 bytes thanks to "regex stealing" by pajonk

Edit2: +21 bytes to remove case-sensitive duplicates, while returning the originally cased input

q=(o=grep("^s(?!h)",scan(,""),T,T,T))[!duplicated(tolower(o))];if(q[7]>F)q[1:7]

Try it online!

Outputs the first 7 unique words starting with 's' or 'S' but not 'sh' or 'Sh', if there are at least 7.
Otherwise errors without outputting anything.

---

R, 59 63 61 82 bytes

(q=(o=grep("^s(?!h)",scan(,""),T,T,T))[!duplicated(tolower(o))])[1:7][length(q)>6]

Try it online!

As above, but exits quietly without erroring if there are less than 7 valid words.

Answer 3

JavaScript (ES6),  62  61 bytes

Saved 1 byte thanks to @l4m2

s=>(a=[...new Set(s.match(/\bs(?!h)\w*/g))]).slice(a[6]||7,7)

Try it online!

Answer 4

Retina 0.8.2, 47 bytes

+msi`(^(.+)$.+)^\2$
$1
Gi`^s(?!h)
1!`.+(¶.+){6}

Try it online! Explanation:

+msi`(^(.+)$.+)^\2$
$1

Delete case-insensitive duplicates.

Gi`^s(?!h)

Keep only words beginning with s but not sh.

1!`.+(¶.+){6}

Select the first seven words.

Answer 5

Jelly, 24 bytes

^{Assuming that we must handle uppercase S and H too and that we must/may return the leftmost distinct "s-words"}

ḣ2ŒliⱮ⁼Ø.
ḲQçƇ⁾hsḣJf7ḢƊ$

A monadic Link that accepts a list of characters and yields a list of the words.

Try it online!

How?

ḣ2ŒliⱮ⁼Ø. - Helper Link, valid word?: list of characters, Word; identifiers ("hs")
ḣ2        - head Word to index two   - e.g. "Child" -> "Ch"
  Œl      - lower-case -> X                     -> X = "ch"
     Ɱ    - map across C in identifiers with:
    i     -   first (1-indexed) index of C in X     -> [2,0]
                                                    ('h' at index 2, no 's' exists)
       Ø. - [0,1]
      ⁼   - equal?

ḲQçƇ⁾hsḣJf7ḢƊ$ - Link get s-words: list of characters, T
Ḳ              - split T at space characters -> Words
 Q             - deduplicate
    ⁾hs        - set the right argument to "hs"
   Ƈ           - filter keep those Words for which:
  ç            -   call the helper Link as a dyad - f(Word, "hs")
             $ - last two links as a monad - f(ValidWords):
            Ɗ  -   last three links as a monad - g(ValidWords):
        J      -     range of length -> [1,2,...,number of valid words]
          7    -     seven
         f     -     filter-keep -> [7] or [] if less than seven valid words
           Ḣ   -     head        -> 7   or 0
       ḣ       -   head of ValidWords to that index

Answer 6

05AB1E, 27 (or 22?) bytes

#ʒlć'sQsн'hÊ*}DlDÙkèDg7@7*£

Assumes differently cased words (e.g. speed/SPEED/sPeEd) are all the same for the uniquify. Otherwise this could have been 22 bytes by replacing the DlDÙkè with Ù.

Try it online or verify some more test cases.

Explanation:

#          # Split the (implicit) input-string by spaces
 ʒ         # Filter this list of words by:
  l        #  Convert it to lowercase
   ć       #  Extract head; pop remainder-string and first char separated
    'sQ   '#  Check if this head is an "s"
   s       #  Swap so the remainder-string is at the top
    н      #  Pop and push its first character
     'hÊ  '#  Check that it's NOT equal to a "h"
   *       #  Check that both were truthy
 }D        # After the filter: duplicate the resulting list of words
   l       # Convert each to lowercase
    D      # Duplicate it again
     Ù     # Uniquify the top copy
      k    # Get all its indices in the lowercase list
       è   # Use it to index in the regular case-insensitive list
   D       # Duplicate the list
    g      # Pop and push its length
     7@    # Check if it's >=7
       7*  # Multiply that 0/1 by 7 (either 0 or 7)
         £ # Leave that many leading words from the list
           # (after which the resulting list is output implicitly)

Answer 7

Perl 5, 60 bytes

my%s;$_=(@r=grep!$s{lc$_}++,/\bs(?!h)\w*/gi)<7?"":"@r[0..6]"

Try it online!

Answer 8

Python 3, 111 102 103 bytes

I am pretty sure this isn't a perfect solution, as my regex skills are far from perfect and this seems an unnecessarily long way to check for "any word character that is not h or H", but it works. Takes a list of the words, and returns a set of the seven words, or nothing if it cannot find seven words.

import re
def f(x):
 r=set()
 for i in x:
  if re.match("[Ss](?![Hh])",i):r|={i}
  if len(r)>6:return r

Edit -9 bytes: realized that it didn't have to check if the words only contained alphanumerics

Edit +1 bytes: @a stone arachnid pointed out my regex failed for input of s

Try it online!

Answer 9

PowerShell Core for Windows, 68 bytes

Thanks @Julian for the inspiration

($r=$args|sls '\bs(?!h)\S*'-a|% m*|% v*e|sort -u -t 7)*!($r.count-7)

Try it online!

The alias sort is not defined for Linux PowerShell and TIO. Linux requires sort-object.

Less golfed:

$result = $args|select-string '\bs(?!h)\S*' -allMatches|% matches|% value|sort -unique -top 7
$result*($result.count-eq7)

Answer 10

PowerShell Core, 78 75 bytes

($r=-split$args-match'^s[^h]*$'|?{!($_-in$u);$u+=,$_})[0..6]*($r.count-ge7)

Try it online!

-3 bytes thanks to mazzy!

Answer 11

Charcoal, 38 bytes

ＷＳ⊞υι≔Φυ∧›⁼↧§ι⁰s⁼↧§ι¹h⁼κ⌕↧υ↧ιυ…υ∧‹⁶Ｌυ⁷

Try it online! Link is to verbose version of code. Takes input as a list of newline-terminated strings. Explanation:

ＷＳ⊞υι

Input the list of words.

≔Φυ∧›⁼↧§ι⁰s⁼↧§ι¹h⁼κ⌕↧υ↧ιυ

Case-insensitively filter out all words that don't start with an s or start with sh or are duplicate.

…υ∧‹⁶Ｌυ⁷

Output the first seven remaining words if there are more than six of them.

Answer 12

Japt, 21 bytes

I think this is right; it's been a long day so I'm not at all sure!

¸fÈÎos «XÅÎohÃòk mί7

Try it

Answer 13

JavaScript (Node.js), 64 bytes

s=>(a=s.match(/(\bs(?!h)\w*\b)(?<!\b\1 .*)/gi)).slice(a[6]||7,7)

Try it online!

Fixed Arnauld's uppercase problem with 3 bytes

Answer 14

PHP, 134 131 bytes

$a=explode(" ",$argn);foreach($a as$v){$l=strtolower($v);if($l[0]=="s"&&$l[1]<>"h"){$b[$l]="$v ";if(count($b)==7){echo join($b);}}}

Try it online!

Explanation: Used associative array key to prevent dups instead of in_array function.

----- Previous answer -----

$a=explode(" ",$argn);foreach($a as$v){if($v[0]=="s"&&$v[1]<>"h"&&!in_array($v,$b)&&$i<7){$b[]=$v;$i++;}}if($i==7){echo join($b," ");}

Try it online!

Explanation: Straightforward conversion of input to array, step through array testing each word against the rules, if a word passes the test add it to second array, and only print second array if 7 words pass the test.