Sort letters by height
Given a sequence of lower case letters, sort them into order of height.
Heights
The heights of letters are dependent on font, so for this challenge the height order to be used is as defined below:
acemnorsuvwxz
t
i
bdfghklpqy
j
Letters on the same line are defined to be the same height. The first line is the shortest letters, the last line is the tallest letter.
Input
- A sequence of lower case letters
- This may be a string or any data structure of characters
Output
- A sequence of the same letters in sorted order
- This may be a string or any ordered data structure of characters. It does not need to match the input format (provided it is consistent between inputs)
- For example, you may take input as an array of characters, and output as a string, provided this format does not change for different inputs
- The sort does not need to be stable (letters that are the same height do not need to remain in the same order as the input, even if the input was already sorted)
- The sort may be either ascending or descending
Test cases
Since the sort order can be either ascending or descending, and the sort does not need to be stable, most inputs will have many possible valid outputs. The test cases are in the format "input" : ["valid", "outputs"]
. You may choose any of the valid outputs, but you must output only one of them.
"a" : ["a"]
"aa" : ["aa"]
"atibj" : ["atibj", "jbita"]
"tick" : ["ctik", "kitc"]
"now" : ["now", "nwo", "onw", "own", "wno", "won"]
"just" : ["jtus", "jtsu", "sutj", "ustj"]
"pjztyix" : ["xztipyj", "zxtipyj", "xztiypj", "zxtiypj", "jypitzx", "jypitxz", "jpyitzx", "jpyitxz"]
Explanations are optional, but I'm more likely to upvote answers that have one.
Vyxal, 12 bytes ``` µ«,←⋎„¶ɽ …
2y ago
Ruby, 53 51 bytes ```ruby …
1y ago
[Python 3], 43 bytes …
2y ago
Japt, 23 16 15 13 bytes Lex …
1y ago
[C (gcc)], 152 bytes …
2y ago
5 answers
C (gcc), 152 bytes
r;e(c){r=c=='j'?5:strchr("bdfghklpqy",c)?4:c=='i'?3:c=='t'?2:1;}c;s;f(char*i,char*o){s?(e(*i),c=r,e(*o)):(qsort(i,strlen(i),s=1,f),puts(i));return r-c;}
Output:
a
aa
jbita
kitc
now
jtus
jpyitzx
I didn't fine tune it much, but it's delightfully crazy :)
Explanation:
- The function
f
takes one input string and one output string as parameter. It returnsint
as per gcc's old C90 default behavior. - Since in-place modification of the input should be fine according to Codidacts input/output rules, I actually never use the output but modify the input. (Would crash & burn in case of string literals though, but the challenge only said strings, so I skipped the copy from input to output buffer.)
- The global variable
s
gets zero-initialized and is used to mark the function's state. If not set, then it's the first call. If set, then the function has been called before. (This will have to be reset to zero externally, in order to execute multiple test cases in a row from main()) -
s?
checks ifs
is zero, if so executeqsort
and then print. -
qsort
uses the very same functionf
as its own callback to save a function. This abuses the fact thatconst void*
andchar*
likely have the same representation. And the function returnsint
so it fulfils the requirements of a qsort callback. Curiously,qsort
didn't raise a compiler error when passed the wrong type of function pointer (this doesn't sound conforming at all, gcc...). - When
qsort
is called, the size of one item parameter also sets the states
to 1. -
f
gets called byqsort
many times and now thes?
evaluates to 1. It calls the helper functione
to evaluate both parameters (i
ando
are now the qsort callback parameters). -
e
sets a global variabler
to value 5,4,3,2 or 1 depending on how which set of characters the passed character belongs to. - No benefit from typing out decimal values of these characters since they are all above ASCII value 100, might as well use
'a'
character constants then. - The result
r
from the first call toe
is stored inc
. Then after the next callr-c
is returned as the result of qsort callback. This weighs'j'
as lowest, swap toc-r
to get the other way around. - The function actually always returns
r-c
but whens==0
nobody cares since the result is placed in thei
buffer anyway. - When
qsort
is done, we are back to the first execution of the function, soputs
is called once before finishing.
0 comment threads
Vyxal, 12 bytes
µ«,←⋎„¶ɽ₌Ż«ḟ
Outputs as a list of characters. Sort by index in (compressed) tibdfghklpqyj
.
0 comment threads
Ruby, 53 51 bytes
->i{i.chars.sort_by{"tibdfghklpqyj".index(_1)||-1}}
Works in Ruby 2.7 and Ruby 3.
Explanation
-
->i{...}
is a short way to define a 1-argument lambda with parameter i -
.chars
will turn a string into an array of characters -
.sort_by
will do a normal sort, transforming each element for sorting purposes to whatever the block returns. -
_1
tells ruby that my block actually wanted a variable and that I'm using it at that location. -
a.index(b)
returns the index ofa
inb
. It returnsnil
when not found. -
a||-1
ifa
is falsy, make it-1
instead. (And falsy is much better defined than in JavaScript, this will only turnnil
andfalse
into -1)
Fun facts
In Ruby parentheses are optional, so I would have almost been able to do .index _1||-1
. Unfortunately, the precedence of ||
is such that this now works on the argument (_1
) rather than the output of the whole function. One could use or
since it has different precedence than ||
, but this doesn't save space because you now need an extra space. So all this yields us an equally long but possibly more convoluted:
->i{i.chars.sort_by{"tibdfghklpqyj".index _1 or-1}}
Python 3, 43 bytes
lambda x:sorted(x,key='tibdfghklpqyj'.find)
Sorts based on index in a sorted string. Inputs as a string and outputs as a list.
-14 bytes by replacing .index()
with .find()
0 comment threads