# Split a byte array into a bit array

Write a function that when given a buffer b (1 - 104857600 bytes long) and a number of bits n (1 <= n <= 64), splits the buffer into chunks of n bits. Right-pad the last chunk with 0s up to n bits.

e.g.

Given the buffer b = "f0oBaR" or equivalently [102,48,111,66,97,82] and n = 5, return

[12, 24, 24, 6, 30, 16, 19, 1, 10, 8]


This is because the above buffer, when represented as binary looks like:

01100110 00110000 01101111 01000010 01100001 01010010


And when re-grouped into 5s looks like:

01100 11000 11000 00110 11110 10000 10011 00001 01010 010[00]


Which when converted back into decimal gives the answer.

## Notes

• You may use whatever data type makes the most sense in your language to represent the buffer. In PHP you'd probably use a string, in Node you might want to use a Buffer
• If you use a string to represent the buffer, assume it's ASCII for the char -> int conversion
• You may use an array of ints (0-255) for input if you prefer
• Return value must be an array or list of ints

## Test Cases

> b = "Hello World", n = 50
318401791769729, 412278856237056

> b = [1,2,3,4,5], n = 1
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1

> b = "codegolf", n = 32
1668244581, 1735355494

> b = "codegolf" n = 64
7165055918859578470

> b = "codegolf" n = 7
49, 91, 108, 70, 43, 29, 94, 108, 51, 0

> b = "Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque vel est eu velit lacinia iaculis. Nulla facilisi. Mauris vitae elit sapien. Nullam odio nulla, laoreet at lorem eu, elementum ultricies libero. Praesent orci elit, sodales consectetur magna eget, pulvinar eleifend mi. Ut euismod leo ut tortor ultrices blandit. Praesent dapibus tincidunt velit vitae viverra. Nam posuere dui quis ipsum iaculis, quis tristique nisl tincidunt. Aliquam ac ligula a diam congue tempus sit amet quis nisl. Nam lacinia ante vitae leo efficitur, eu tincidunt metus condimentum. Cras euismod quis quam vitae imperdiet. Ut at est turpis.", n = 16
19567, 29285, 27936, 26992, 29557, 27936, 25711, 27759, 29216, 29545, 29728, 24941, 25972, 11296, 25455, 28275, 25955, 29797, 29813, 29216, 24932, 26992, 26995, 25449, 28263, 8293, 27753, 29742, 8272, 25964, 27749, 28276, 25971, 29045, 25888, 30309, 27680, 25971, 29728, 25973, 8310, 25964, 26996, 8300, 24931, 26990, 26977, 8297, 24931, 30060, 26995, 11808, 20085, 27756, 24864, 26209, 25449, 27753, 29545, 11808, 19809, 30066, 26995, 8310, 26996, 24933, 8293, 27753, 29728, 29537, 28777, 25966, 11808, 20085, 27756, 24941, 8303, 25705, 28448, 28277, 27756, 24876, 8300, 24943, 29285, 25972, 8289, 29728, 27759, 29285, 27936, 25973, 11296, 25964, 25965, 25966, 29813, 27936, 30060, 29810, 26979, 26981, 29472, 27753, 25189, 29295, 11808, 20594, 24933, 29541, 28276, 8303, 29283, 26912, 25964, 26996, 11296, 29551, 25697, 27749, 29472, 25455, 28275, 25955, 29797, 29813, 29216, 28001, 26478, 24864, 25959, 25972, 11296, 28789, 27766, 26990, 24946, 8293, 27749, 26982, 25966, 25632, 28009, 11808, 21876, 8293, 30057, 29549, 28516, 8300, 25967, 8309, 29728, 29807, 29300, 28530, 8309, 27764, 29289, 25445, 29472, 25196, 24942, 25705, 29742, 8272, 29281, 25971, 25966, 29728, 25697, 28777, 25205, 29472, 29801, 28259, 26980, 30062, 29728, 30309, 27753, 29728, 30313, 29793, 25888, 30313, 30309, 29298, 24878, 8270, 24941, 8304, 28531, 30053, 29285, 8292, 30057, 8305, 30057, 29472, 26992, 29557, 27936, 26977, 25461, 27753, 29484, 8305, 30057, 29472, 29810, 26995, 29801, 29045, 25888, 28265, 29548, 8308, 26990, 25449, 25717, 28276, 11808, 16748, 26993, 30049, 27936, 24931, 8300, 26983, 30060, 24864, 24864, 25705, 24941, 8291, 28526, 26485, 25888, 29797, 28016, 30067, 8307, 26996, 8289, 28005, 29728, 29045, 26995, 8302, 26995, 27694, 8270, 24941, 8300, 24931, 26990, 26977, 8289, 28276, 25888, 30313, 29793, 25888, 27749, 28448, 25958, 26217, 25449, 29813, 29228, 8293, 29984, 29801, 28259, 26980, 30062, 29728, 28005, 29813, 29472, 25455, 28260, 26989, 25966, 29813, 27950, 8259, 29281, 29472, 25973, 26995, 28015, 25632, 29045, 26995, 8305, 30049, 27936, 30313, 29793, 25888, 26989, 28773, 29284, 26981, 29742, 8277, 29728, 24948, 8293, 29556, 8308, 30066, 28777, 29486

> b = [2,31,73,127,179,233], n = 8
2, 31, 73, 127, 179, 233

• Is it supposed to work for values of n greater than 8? If so, what about values of n greater than 64, which is larger than most language's integer precision. – speedplane Mar 22 '16 at 5:23
• Why does the return value have to be ints? – wizzwizz4 Mar 22 '16 at 7:19
• @wizzwizz4 I don't think so. They can't be bytes because they don't have 8 bits. Bitwise operators normally work on ints and not much else. If you have a better suggestion then I'm listening, but otherwise ints it is. – mpen Mar 22 '16 at 20:22
• @wizzwizz4 Because I don't want people to be able to skip a step. I don't want answers like "the first 5 bits of this byte contain the answer" -- the result should not contain any superfluous information, and it should be easily converted back to ASCII or some character mapping (a real-life use-case). Also, given the number of answers so far, it doesn't appear to be a problem. – mpen Mar 22 '16 at 21:40
• @mpen I see the confusion. char is an integer that happens to be one byte long. – wizzwizz4 Mar 23 '16 at 17:34

## Pyth, 18 17 bytes

iR2c.[t.B+C1z\0QQ


Thanks to @lirtosiast for a byte!

            z      get input
+C1       prepend a 0x01 to prevent leading zeroes from disappearing
.B          convert to binary string
t            remove the leading 1 from ^^
.[       \0Q   pad right with zeroes to multiple of second input
c            Q  get chunks/slices of length second input
iR2                map(x: int(x, 2))


# Jelly, 13 bytes

1;ḅ256æ«BḊsḄṖ


This takes the input as a list of integers. Try it online!

### How it works

1;ḅ256æ«BḊsḄṖ  Main link. Arguments: A (list), n (integer)

1;             Prepend 1 to A.
ḅ256         Convert from base 256 to integer.
æ«       Bitshift the result n units to the left.
B      Convert to binary.
Ḋ     Discard the first binary digit (corresponds to prepended 1).
s    Split into chunks of length n.
Ḅ   Convert each chunk from binary to integer.


# Julia, 117 bytes

f(x,n,b=join(map(i->bin(i,8),x)),d=endof,z=rpad(b,d(b)+d(b)%n,0))=map(i->parse(Int,i,2),[z[i:i+n-1]for i=1:n:d(z)-n])


This is a function that accepts an integer array and an integer and returns an integer array. It's an exercise in function argument abuse.

Ungolfed:

function f(x::Array{Int,1},                  # Input array
n::Int,                           # Input integer
b = join(map(i -> bin(i, 8), x)), # x joined as a binary string
d = endof,                        # Store the endof function
z = rpad(b, d(b) + d(b) % n, 0))  # b padded to a multiple of n

# Parse out the integers in base 2
map(i -> parse(Int, i, 2), [z[i:i+n-1] for i = 1:n:d(z)-n])
end

• Why did you temporarily delete it? – CalculatorFeline Mar 22 '16 at 14:25
• @CatsAreFluffy I realized I had done something wrong initially such that it worked for the test case but wouldn't necessarily in general. Should be all good now though. :) – Alex A. Mar 22 '16 at 17:06

## JavaScript (ES6), 120 bytes

f=(a,n,b=0,t=0,r=[])=>b<n?a.length?f(a.slice(1),n,b+8,t*256+a[0],r):b?[...r,t<<n-b]:r:f(a,n,b-=n,t&(1<<b)-1,[...r,t>>b])


Recursive bit twiddling on integer arrays. Ungolfed:

function bits(array, nbits) {
var count = 0;
var total = 0;
var result = [];
for (;;) {
if (nbits <= count) {
// We have enough bits to be able to add to the result
count -= nbits;
result.push(total >> count);
total &= (1 << count) - 1;
} else if (array.length) {
// Grab the next 8 bits from the array element
count += 8;
total <<= 8;
total += array.shift();
} else {
// Deal with any leftover bits
if (count) result.push(total << nbits - count);
return result;
}
}
}

• @WashingtonGuedes I managed to golf another 9 bytes off my own golf of your solution, but it's still 129 bytes, sorry: "(s,n)=>(s.replace(/./g,x=>(256+x.charCodeAt()).toString(2).slice(1))+'0'.repeat(n-1)).match(eval(/.{${n}}/g)).map(x=>+0b${x})".length – Neil Mar 22 '16 at 9:57
• You sure this one runs? The ungolfed version is crashing Chrome. – mpen Mar 22 '16 at 16:20
• @mpen The golfed version definitely runs on Firefox. The ungolfed version may have errors in it. – Neil Mar 22 '16 at 16:34
• Aha! And so it does. I thought Chrome's JS engine was ahead FF's but I guess not. – mpen Mar 22 '16 at 20:26
• @mpen Fixed a couple of subtle bugs in my ungolfed code for you. – Neil Mar 22 '16 at 21:45

## Python 3, 102 bytes

j=''.join
lambda s,n:[int(j(k),2)for k in zip(*[iter(j([bin(i)[2:].zfill(8)for i in s+[0]]))]*n)][:-1]


use iter trick to group string

• s: the input string/buffer
• n: the number of bits in each splitted chunk

## Results

>>> f([102,48,111,66,97,82],4)
[6, 6, 3, 0, 6, 15, 4, 2, 6, 1, 5, 2, 0]

>>> f([102,48,111,66,97,82],5)
[12, 24, 24, 6, 30, 16, 19, 1, 10, 8]

>>> f([102,48,111,66,97,82],6)
[25, 35, 1, 47, 16, 38, 5, 18]

>>> f([102,48,111,66,97,82],8)
[102, 48, 111, 66, 97, 82]

• You should've documented your code to explain what does the 'n' param do? – Quadcubic Dec 8 '19 at 5:58
• @nullptr n have the same meaning than the question : splits the buffer into chunks of n bits – Erwan Dec 10 '19 at 8:03

# Ruby, 114 bytes

->s,n{a=s.bytes.map{|b|b.to_s(2).rjust 8,?0}.join.split""
r=[]
r<<a.shift(n).join.ljust(n,?0).to_i(2)while a[0]
r}


Slightly cleaner:

f = -> str, num {
arr = str.bytes.map {|byte|
byte.to_s(2).rjust(8, "0")
}.join.split("")
result = []
while arr.size > 0
result << arr.shift(num).join.ljust(num, "0").to_i(2)
end
result
}

puts f["f0oBaR", 5]


# Perl 6, 93 68 bytes

{@^a».&{sprintf "%08b",$_}.join.comb($^b)».&{:2($_~0 x$b-.chars)}}


## PHP, 262217189 bytes

function f($b,$n){$M='array_map';return$M('bindec',$M(function($x)use($n){return str_pad($x,$n,0);},str_split(implode('',$M(function($s){return str_pad($s,8,0,0);},$M('decbin',$b))),$n)));}  (updated with tips from Ismael Miguel) Formatted for readability: function f($b, $n) {$M = 'array_map';
return $M('bindec',$M(function ($x) use ($n) {
return str_pad($x,$n, 0);
}, str_split(implode('', $M(function ($s) {
return str_pad($s, 8, 0, 0); },$M('decbin', $b))),$n)));
}


Example:

> implode(', ',f(array_map('ord',str_split('f0oBaR')),5));
"12, 24, 24, 6, 30, 16, 19, 1, 10, 8"

• Instead of str_pad($s,8,'0',STR_PAD_LEFT), you can use str_pad($s,8,0,0). You can remove the quotes on bindec and decbin to save 4 bytes. To save more, you can store array_map in a variable and pass it instead. Here you go: function f($b,$n){$M=array_map;return$M(bindec,$M(function($x)use($n){return str_pad($x,$n,0);},str_split($M('',array_map(function($s){return str_pad($s,8,0,0);},$M(decbin,$b))),5)));} (184 bytes). – Ismael Miguel Mar 22 '16 at 9:30
+ '\n'+ r.map(x=><i>${(256*256*256*256+x).toString(2).slice(-g)}</i>).join } test() #A { width: 50% } #G { width: 5% } i:nth-child(even) { color: #00c } i:nth-child(odd) { color: #c00 } Input array <input id=A value="102,48,111,66,97,82"> Group by bits <input id=G value=5> (up to 32)<br> Output <button onclick="test()">-></button> <span id=O></span> <pre id=K></pre> • Instead of doubling x each time, why not shift x right i bits? – Neil Mar 22 '16 at 21:34 • @Neil uh...why...idiocy? – edc65 Mar 22 '16 at 22:23 • I just noticed that c-g?[...s,t<<c]:s might save you a couple more bytes. – Neil Mar 23 '16 at 0:47 • @Neil this requires some thoughts – edc65 Mar 23 '16 at 8:56 # J, 24 bytes [:#.-@[>\;@(_8:{."1#:@])  This is an anonymous function, which takes n as its left argument and b as numbers as its right argument. Test:  5 ([:#.-@[>\;@(_8:{."1#:@])) 102 48 111 66 97 82 12 24 24 6 30 16 19 1 10 8  Explanation: [:#.-@[>\;@(_8:{."1#:@]) #:@] NB. Convert each number in b to bits _8:{."1 NB. Take the last 8 items for each NB. (padding with zeroes at the front) ;@ NB. Make a list of all the bits -@[ NB. Negate n NB. (\ gives non-overlapping infixes if [<0) >\ NB. Get non-overlapping n-sized infixes [:#. NB. Convert those back to decimal  ## Haskell, 112 109 bytes import Data.Digits import Data.Lists n#x=unDigits 2.take n.(++[0,0..])<$>chunksOf n(tail.digits 2.(+256)=<<x)


Usage example: 5 # [102,48,111,66,97,82] -> [12,24,24,6,30,16,19,1,10,8].

How it works

import Data.Digits                  -- needed for base 2 conversion
import Data.Lists                   -- needed for "chunksOf", i.e. splitting in
-- sublists of length n

(                  =<<x) -- map over the input list and combine the
-- results into a single list:
tail.digits 2.(+256)    -- convert to base two with exactly 8 digits
chunksOf n                 -- split into chunks of length n
<$> -- convert every chunk (<$> is map)
take n.(++[0,0..])              -- pad with 0s
unDigits 2                          -- convert from base 2


## Java, 313306 322 bytes

I hope this beats PHP... And nope. Stupid long function names.

-7 thanks to @quartata for getting rid of public +16 to fix an error when the split was exact, thanks to @TheCoder for catching it

int[] f(String b,int s){int i=0,o[]=new int[(int)Math.ceil(b.length()*8.0/s)],a=0;String x="",t;for(char c:b.toCharArray()){t=Integer.toString(c,2);while(t.length()<8)t="0"+t;x+=t;a+=8;while(a>=s){o[i++]=Integer.parseInt(x.substring(0,s),2);x=x.substring(s,a);a-=s;}}while(a++<s)x+="0";o[i]=Integer.parseInt(x,2);return o;}

• I don't think you have to make the function public. – a spaghetto Mar 22 '16 at 14:33
• What version of Java did you run this in? It doesn't seem to compile: ideone.com/3tonJt – mpen Mar 22 '16 at 21:44
• @mpen Ah, whoops. I forgot, I changed it on my computer before posting. Will fix. – Blue Mar 23 '16 at 1:41
• @JackAmmo yup, sure did. Stupid tiny phone keyboard. – Blue Mar 23 '16 at 1:50
• o[]=new int[b.length()*8/s+1] - This will assign wrong size If (b.length()*8)%s==0 – The Coder Mar 24 '16 at 10:18

# Ruby, 66 bytes

->s,n{(s.unpack('B*')[0]+?0*~-n).scan(/.{#{n}}/).map{|x|x.to_i 2}}


Try it online!

Takes input buffer as a string, so that a few test strings were constructed directly in the footer to avoid unprintables.

# MATL, 9 bytes

8&B!we!XB


Try it online!

Takes input b as a string delimited by '' or as an array of comma-separated values like [102, 48, 111], then n.

8           # push 8
&B          # implicitly take input b, and use 2-element convert to binary
# to push a binary matrix of 8 bits
!           # transpose, so each column represents an input
w           # implicitly take input n and swap it with binary matrix to top of stack
e           # reshape into n rows, padding with zeros at end
# this matrix will have each column as an n-bit integer
!           # transpose, so each row is now the n-bit integer
XB          # convert each row to decimal
# implicit output


# Perl 5-nl -MData::Dump=pp , 96 bytes

$}=$_;pp map{$_=sprintf"%-$}s",$_;y/ /0/;oct"0b$_"}(join'',map{sprintf"%08b",$_}<>)=~m/.{1,$_}/g


Try it online!

Requires the Data::Dump module.

Takes n on the first line of input and the numbers on each line after that.

Outputs to STDERR (the Debug field on TIO).

Deparsed and tidied:

BEGIN { $/ = "\n";$\ = "\n"; }
use Data::Dump ( split( /,/, 'pp', 0 ) );
LINE: while ( defined( $_ = readline ARGV ) ) { chomp$_;
$} =$_;
pp(
map( {
$_ = sprintf( "%-$}s", $_ ); tr/ /0/; oct "0b$_";
} join( '', map( { sprintf '%08b', $_; } readline ARGV ) ) =~ /.{1,$_}/g )
);
}


# Powershell 146 bytes

param([int[]][char[]]$b,$n)-join($b|%{[convert]::ToString($_,2).PadLeft(8,"0")})-split"(.{$n})"|?{$_}|%{[convert]::ToInt32($_.PadRight($n,"0"),2)}


Take in the buffer and convert it to a char array and then an integer array. For each of those convert to binary, pad the entries with 0's where needed, and join as one large string. Split that string on n characters and drop the blanks that are created. Each element from the split is padded (only the last element really would need it) and converted back into an integer. Output is an array

# Python 3.5 - 312 292 bytes:

def d(a, b):
o=[];o+=([str(bin(g)).lstrip('0b')if str(type(g))=="<class 'int'>"else str(bin(ord(g))).lstrip('0b')for g in a]);n=[''.join(o)[i:i+b]for i in range(0,len(''.join(o)),b)];v=[]
for t in n:
if len(t)!=b:n[n.index(t)]=str(t)+'0'*(b-len(t))
v+=([int(str(f),2)for f in n])
return v


Although this may be long, this is, in my knowledge, is the shortest way to accept both functions and arrays without errors, and still being able to retain some accuracy in Python 3.5.

## Java, 253 247 bytes

Golfed

int i,l,a[];Integer I;String f="";int[]c(String s,int n){for(char c:s.toCharArray())f+=f.format("%08d",I.parseInt(I.toString(c, 2)));while(((l=f.length())%n)>0)f+="0";for(a=new int[l=l/n];i<l;)a[i]=I.parseInt(f.substring(i*n,i++*n+n),2);return a;}


UnGolfed

int i,l,a[];
Integer I;
String f="";
int[]c(String s,int n) {
for(char c:s.toCharArray())
f+=f.format("%08d",I.parseInt(I.toString(c,2)));
while(((l=f.length())%n)>0)
f+="0";
for(a=new int[l=l/n];i<l;)
a[i]=I.parseInt(f.substring(i*n,i++*n+n),2);
return a;
}

• c, 2 => c,2; ((l=f.length())%n)>0 => (l=f.length())%n>0; – Zacharý Oct 9 '18 at 20:01

# Jelly, 13 bytes

+256BḊ€Ẏsz0ZḄ


Try it online!

Note: The input is actually a list of non-negative integers, but the TIO link eases the pain for you, and accepts either such a list or a string.

# Stax, 12 bytes

è■àåk┘K¥xk└╣


Run and debug it

This isn't a function as specified in the challenge, but a program, since stax doesn't support functions. It supports input of strings or array literals.

# Python 2, 101 bytes

lambda a,n:[int(''.join(bin(x+256)[3:]for x in a+[0]*n)[n*i:][:n],2)for i in range((len(a)*8-1)/n+1)]


Try it online!

# Dyalog APL, 36 bytes

{2∘⊥¨↓A⍴(×/A←⍺,⍨⌈⍺÷⍨8×≢⍵)↑∊↓⍉⍵⊤⍨8/2}


Try it online!

This probably could be golfed down more.

# Elixir, 63 60 bytes

&(s=&2-1)&&for<<x::size(&2)<-<<"#{&1}",0::size(s)>> >>,do: x


Try it online!

Takes input as Elixir binary, outputs a list of integers.

This code makes use of Elixir bitstring generator comprehension to chunk input binary &1 into bit blocks of size provided as argument &2. To account for any leftover bits at the end, we pad the binary with &2 - 1 zero bits. Here is also the place where some unwanted verbosity kicks in: Elixir complains if we don't explicitly declare &1 as bitstring, and it also doesn't support expressions in size(...), hence the need for an extra variable assignment.

Easter egg: in the footer, replace IO.inspect with IO.puts, and our function magically "translates" Lorem ipsum from Latin to Chinese - Try it online!

# Japt, 16 bytes

After a long day of meetings, it feels like I've forgotten how to golf! Will play around with this on the train home later, see if I can improve on it.

c_¤ùT8ÃòV úTV mÍ


Try it

• Are you saying it's 16 bytes if you encode it with IEC 8859-1? Is that even legal? Did you find a loophole? Did they not specify it has to be UTF-8? Because it's 22 bytes of UTF-8. – mpen Dec 10 '19 at 19:35
• @mpen 1. Each language is free to use it's own encoding/codepage. 2. ISO-8859-1 is a fairly standard encoding used by many languages and not particular to Japt, or this specific solution. 3. Your spec doesn't state that we must count in UTF-8. 4. If it did, you would have been strongly discouraged from it. – Shaggy Dec 10 '19 at 20:46
• I'm not trying to impose anything, I just wanted to know if this was generally accepted, and it appears you're quite right codegolf.meta.stackexchange.com/a/17800/23090 TIO scores Japt the same way, that might be an easier way to qualify it tio.run/##y0osKPn/Pzn@0JLDO0MsDjcf3hSmcHhXSJhC7uHe//8B edit 3: oh your link is pretty much TIO. – mpen Dec 10 '19 at 21:49

# PHP, 135129 124 bytes

function($b,$n){foreach($b as$c)$a.=sprintf('%08b',$c);foreach(str_split($a,$n)as$s)$d[]=bindec(str_pad($s,$n,0));return\$d;}


Try it online!

Implemented as a function, input buffer is an array of ints and returns an array of ints.

Output

> b = "f0oBaR", n = 5
[12,24,24,6,30,16,19,1,10,8]

> b = "Hello World", n = 50
[318401791769729,412278856237056]

> b = [1,2,3,4,5], n = 1
[0,0,0,0,0,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0,0,0,1,1,0,0,0,0,0,1,0,0,0,0,0,0,0,1,0,1]


Verify all test cases

• Nice! Quite a bit shorter than mine. – mpen Apr 3 '19 at 22:53

# APL(NARS), 471 chars, 942 bytes

TH←{v←↑⍴⍴⍵⋄v>2:64⋄v=2:32⋄(v=1)∧''≡0↑⍵:20⋄''≡0↑⍵:4⋄v=1:16⋄⍵≢+⍵:8⋄⍵=⌈⍵:2⋄1}
TV←{x←TH¨⍵⋄k←↑x⋄t←↑⍴⍵⋄t=+/x=2:2⋄t=+/x≤2:1⋄(k≤8)∧⍬≡x∼k:k⋄0}
T←{v←↑⍴⍴⍵⋄v>2:64+TV⍵⋄v=2:32+TV⍵⋄(v=1)∧''≡0↑⍵:20⋄''≡0↑⍵:4⋄v=1:16+TV⍵⋄⍵≢+⍵:8⋄⍵=⌈⍵:2⋄1}
RI←{t←T⍵⋄(t≠1)∧(t≠2)∧(t≠17)∧(t≠18):0⋄∧/((1⊃⍺)≤⍵)∧⍵≤(2⊃⍺)}
B←{(8⍴2)⊤⍵}⋄C←{¯1+⎕AV⍳⍵}⋄f←{t←T⍵⋄(0 255 RI⍵)∧18=t:∊B¨⍵⋄(0 255 RI x←C¨⍵)∧20=t:∊B¨x⋄,¯1}⋄W←{((↑⍴⍵)⍴2)⊥⍵}
q←{(∼1 64 RI,⍺)∨2≠T⍺:,¯1⋄x←f⍵⋄¯1=↑x:,¯1⋄t←↑⍴x⋄k←(⍺-m)×0≠m←⍺∣t⋄W⍉((t+k)÷⍺)⍺⍴(((t⍴1),k⍴0)\x)}


commented code and test:

  ⍝TH⍵ return type its argument
TH←{v←↑⍴⍴⍵⋄v>2:64⋄v=2:32⋄(v=1)∧''≡0↑⍵:20⋄''≡0↑⍵:4⋄v=1:16⋄⍵≢+⍵:8⋄⍵=⌈⍵:2⋄1}
⍝ TV⍵ check if type each element of array ⍵ is the same and basic
⍝ (float(int and float too),int,char,complex) and return its number (or 0 if it is not basic)
TV←{x←TH¨⍵⋄k←↑x⋄t←↑⍴⍵⋄t=+/x=2:2⋄t=+/x≤2:1⋄(k≤8)∧⍬≡x∼k:k⋄0}
⍝ T⍵ return the type of ⍵ [it would be ok if ⍵ is not a function]
⍝|1 Float|2 Int|4 Char|8 Complex,Quaternion or Oction|16 List|32 Matrix|64 Tensor
⍝|17 List Float|18 List Int|20 List Char=string|etc
T←{v←↑⍴⍴⍵⋄v>2:64+TV⍵⋄v=2:32+TV⍵⋄(v=1)∧''≡0↑⍵:20⋄''≡0↑⍵:4⋄v=1:16+TV⍵⋄⍵≢+⍵:8⋄⍵=⌈⍵:2⋄1}
⍝ ⍺RI⍵ check if the numeric array ⍵ has elements in [1⊃⍺ 2⊃⍺]; if type is not ok return 0(false)
RI←{t←T⍵⋄(t≠1)∧(t≠2)∧(t≠17)∧(t≠18):0⋄∧/((1⊃⍺)≤⍵)∧⍵≤(2⊃⍺)}

B←{(8⍴2)⊤⍵}   ⍝ from decimal to binary of element 0..255
C←{¯1+⎕AV⍳⍵}   ⍝ return the number of char that is in array AV seems the ascii number
⍝ f⍵ with ⍵ List int element in 0..255 or String with numeric element 0..255
⍝ return the corrispondence disclosed binary array
f←{t←T⍵⋄(0 255 RI⍵)∧18=t:∊B¨⍵⋄(0 255 RI x←C¨⍵)∧20=t:∊B¨x⋄,¯1}
W←{((↑⍴⍵)⍴2)⊥⍵} ⍝ from list of binary digit to decimal
⍝ the function for the exercise
q←{(∼1 64 RI,⍺)∨2≠T⍺:,¯1⋄x←f⍵⋄¯1=↑x:,¯1⋄t←↑⍴x⋄k←(⍺-m)×0≠m←⍺∣t⋄W⍉((t+k)÷⍺)⍺⍴(((t⍴1),k⍴0)\x)}

5 q    'f0oBaR'
12 24 24 6 30 16 19 1 10 8
50 q "Hello World"
318401791769729 412278856237056
1  q 1 2 3 4 5
0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 1 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 1
32 q "codegolf"
1668244581 1735355494
7 q "codegolf"
49 91 108 70 43 29 94 108 51 0
8 q 2 31 73 127 179 233
2 31 73 127 179 233
64 q 2 31 73 127 179 233
1.529217252E17
65 q 2 31 73 127 179 233
¯1
0 q 2 31 73 127 179 233
¯1
23 q '123'
1612057 4194304
23 q '123∞'
¯1
23 q '1' 2 3
¯1
23 q 2 3.3
¯1
23 q 2
¯1
23 q '1'
¯1
23 q ,2
65536
23 q ,'1'
1605632