# Rebuild a rectangular array from a corner

I once had a beautiful rectangular array. It was very symmetrical, but unfortunately it has fallen apart and now I only have the top left corner. Your task will be to rebuild the original array.

Your program will receive a 2 dimensional array of integers. For ease of parsing, you may assume they are all between 1 and 9. Your task is to reverse the array's columns, its rows, and both, stitch together the resulting corners, and return the resulting array.

You can assume the array dimensions will be at least 1x1.

Test cases:

Input:
1 2 3
4 5 6

Output:
1 2 3 3 2 1
4 5 6 6 5 4
4 5 6 6 5 4
1 2 3 3 2 1


Input:
1

Output:
1 1
1 1


Input:
9
9
9

Output:
9 9
9 9
9 9
9 9
9 9
9 9


This is , fewest bytes wins!

• I'll bet charcoal can do this in under 10 – FantaC Mar 5 '18 at 16:19
• @tbfninja chat.stackexchange.com/transcript/message/43184083#43184083 but could maybe be shorter with a different input format. – Pavel Mar 5 '18 at 16:20
• @MagicOctopusUrn yes – Pavel Mar 5 '18 at 16:39
• @tfbninja ＷＳ⟦ι⟧‖Ｍ→↓ perhaps? 5 bytes to read the input and 4 to reflect it. – Neil Mar 5 '18 at 17:51
• I'm 99% sure that there is a lang that do this with ╬ (or some similar character) just can't remember which one :c – Rod Mar 5 '18 at 19:36

# Proton, 29 bytes

a=>[b+b[by-1]for b:a+a[by-1]]


Try it online!

There are a few other interesting approaches though:

# Proton, 29 bytes

a=>map(g,(g=x=>x+x[by-1])(a))


Try it online!

You can define the mirror sub-function g in-line, because Proton. It's not shorter though.

# Proton, 36 bytes

(a=>[xfor x:zip(*(a+a[by-1]))])*2


Try it online!

This should be (a=>zip(*(a+a[by-1])))*2 which is 24 bytes, but the zip function is completely broken. Basically, you mirror it and zip, and then do that twice (you can multiply a function by a positive integer to apply the function multiple times).

# Canvas, 1 byte

╬


Try it here!

Outputs as a multiline string

• Nice, @Rod called that one hah! Is this your language too Dzaima? – Magic Octopus Urn Mar 5 '18 at 20:49
• Damn... I should've remembered... – totallyhuman Mar 5 '18 at 20:53
• @MagicOctopusUrn Yeah, IIRC it's meant to basically be S.O.G.L. II: Electric Boogaloo? – ASCII-only Mar 6 '18 at 10:50

r=(++)<*>reverse
r.map r


Try it online!

# Python 3, 38 bytes

lambda a:[b+b[::-1]for b in a+a[::-1]]


Try it online!

Takes a list of lists and returns a list of lists.

Explanation:

lambda a:                              # anonymous lambda function
for b in a+a[::-1]  # for each row in the array and the upside-down array
b+b[::-1]                    # the row with its reverse appended
[                           ] # return in a list


# Husk,  7  6 bytes

Coincidentally, Erik had posted the exact same code in the Husk chatroom about a minute before I posted this.

‼oTS+↔


Try it online!

Pervious version, 7 bytes:

mS+↔S+↔


## Retina, 13 bytes

\%^$ Vs  Try it online! ### Explanation \%$
$^$


On each line (%), match the end of the line ($), and insert the reverse ($^) of the entire line ($) and print the result with a trailing linefeed (\). This does the reflection along the vertical axis and prints the first half of the output. Vs  This just reverses the entire string, which is equivalent to a 180° degree rotation, or in our case (due to the horizontal symmetry) a reflection along the horizontal axis. This way this works is that V's (reverse) default regex is (?m:^.*$), which normally matches each line of the string. However, we activate the singleline option s, which makes . match linefeeds as well and therefore this default regex actually matches the entire string.

The result of this is printed automatically at the end of the program, giving us the second half of the output.

• This doesn't look like any regex flavor I know of :P – Pavel Mar 5 '18 at 16:36
• @Pavel Because Retina isn't just regex. :) – Erik the Outgolfer Mar 5 '18 at 16:36

# MATL, 5 bytes

,tPv!


Try it online!

Explanation:

(implicit input)
,               # do twice:
t              # dup top of stack
P              # flip vertically
v              # vertically concatenate
!              # transpose
(implicit output)

# Octave,  33  29 bytes

Thanks to @Giuseppe for golfing four bytes!

@(A)[B=[A;flip(A)] fliplr(B)]


Try it online!

# JavaScript (Node.js), 625549 46 bytes

A=>(j=x=>[...x,...[...x].reverse()])(A).map(j)


Try it online!

Because Array.prototype.reverse() reverses the array in place, I have to make a shallow copy somewhere first. A=>(j=x=>[...x,...x.reverse()])(A).map(j) does not work.

# J, 12 bytes

(,|.)@,.|."1


Try it online!

## Explanation

         |."1 - reverse each row
,.     - and stitch them to the input
(   )@       - and
,|.         - append the rows in reversed order


# awk, 88 bytes

{s="";for(i=NF;i>0;i--)s=" "$i s" "$i;a[FNR]=s;print s}END{for(i=NR;i>0;i--)print a[i]}

• Welcome to PPCG! Nice first answer :) – HyperNeutrino Mar 8 '18 at 15:34

# Triangularity, 31 bytes

...)...
..IEM..
.DRs+}.
DRs+...


Try it online!

## Explanation

Removing the characters that make up for the padding, here is what the program does:

)IEMDRs+}DRs+ – Full program. Takes a matrix as a 2D list from STDIN.
)             – Push a 0 onto the stack.
I            – Take the input at that index.
E           – Evaluate it.
M    }     – For each row...
DR        – Duplicate and replace the second copy by its reverse.
s+      – Swap and append.
DR   – Duplicate the result and replace the second copy by its reverse.
s+ – Swap and append.


# R, 57 bytes

function(m)rbind(N<-cbind(m,m[,ncol(m):1]),N[nrow(N):1,])


Try it online!

# APL+WIN, 11 bytes

Prompts for a 2d array of integers.

m⍪⊖m←m,⌽m←⎕


# Stax, 5 bytes

:mm:m


Run and debug it online

:m means mirror, which is input.concat(reverse(input)). m, in this context means output each line after applying...

So, mirror the array of rows, and then mirror each row and output.

# Japt, 6 bytes

mê1 ê1


Try it here

## Explanation

           :Implicit input of 2D array
m          :Map
ê1        :  Mirror sub array
ê1     :Mirror main array


# Mathematica, 29 bytes

(g=#~Join~Reverse@#&)@*Map[g]


Try it online!

# SOGL V0.12, 2 bytes

-1 byte thanks to dzaima.

╬ø


Try it here!

# APL (Dyalog Classic), 7 bytes

⍪∘⊖⍨⊢,⌽


Try it online!

# Ruby, 35 bytes

->a{r=->b{b+b.reverse}
r[a].map &r}


Try it online!

A lambda accepting a 2D array and returning a 2D array. It's straightforward, but here's the ungolfed version anyway:

->a{
r=->b{ b+b.reverse } # r is a lambda that returns the argument and its reverse
r[a].map &r          # Add the array's reverse, then add each row's reverse
}


# Java 8, 140 131 bytes

m->{String r="";for(int a=m.length,b=m.length,i=a+a,j;i-->0;r+="\n")for(j=b+b;j-->0;)r+=m[i<a?i:a+a+~i][j<b?j:b+b+~j];return r;}


Explanation:

Try it online.

m->{                      // Method with integer-matrix parameter and String return-type
String r="";            //  Result-String, starting empty
for(int a=m.length,     //  Amount of rows of the input-matrix
b=m.length,  //  Amount of columns of the input-matrix
i=a+a,j;        //  Index integers
i-->0;              //  Loop over double the rows
r+="\n")            //    After every iteration: append a new-line to the result
for(j=b+b;j-->0;)    //   Inner loop over double the columns
r+=                //    Append the result with:
m[i<a?          //     If i is smaller than the amount of rows
i            //      Use i as index in the input-matrix
:             //     Else:
a+a+~i]      //      Use a+a+i-1 as index instead
[j<b?          //     If j is smaller than the amount of columns
j            //      Use j as index in the input-matrix
:             //     Else:
b+b+~j];     //      Use b+b+j-1 as index instead
return r;}              //  Return the result-String


# J, 11 bytes

Anonymous tacit prefix function.

|:@(,|.)^:2


Try it online!

|: transpose

@(…) the result of:

, the argument followed by

|. its reverse

^:2 and all this done twice

# SNOBOL4 (CSNOBOL4), 119 113 bytes

	T =TABLE()
I	X =X + 1
I =INPUT	:F(D)
OUTPUT =T<X> =I REVERSE(I)	:(I)
D	X =X - 1
OUTPUT =GT(X) T<X>	:S(D)
END


Try it online!

Takes input as strings on STDIN, without spaces. This only works because the digits are 1-9 and would fail otherwise.

• I can see why people don't use this language anymore. This is so weird. – Pavel Mar 5 '18 at 17:15
• @Pavel SNOBOL is truly a terrible language to work with. this is a more modern C implementation of it which has additional builtin functions like REVERSE; the original only supported integer arithmetic as well, as far as I can tell. – Giuseppe Mar 5 '18 at 17:29

# C (gcc), 114 111 bytes

j,i;f(A,w,h)int*A;{for(i=h+h;i-->0;puts(""))for(j=w+w;j-->0;)printf("%d,",A[(i<h?i:h+h+~i)*w+(j<w?j:w+w+~j)]);}


Try it online!

# C (gcc), 109 bytes (abusing ease of parsing)

• Thanks to Kevin Cruijssen for suggesting to only allow one-digit input integers; saved two bytes.
j,i;f(A,w,h)int*A;{for(i=h+h;i-->0;puts(""))for(j=w+w;j-->0;)putchar(A[(i<h?i:h+h+~i)*w+(j<w?j:w+w+~j)]+48);}


Try it online!

• You can save 3 bytes by inverting the loops. for(i=h+h;i-->0;puts(""))for(j=w+w;j-->0;) – Kevin Cruijssen Mar 6 '18 at 8:18
• Does not fulfill spec; prints the array, rather than returning it. – user77406 Mar 6 '18 at 11:14
• "For ease of parsing, you may assume they are all between 1 and 9.", so you can remove the comma in the printf("%d" for an additional -1 byte. – Kevin Cruijssen Mar 6 '18 at 15:12
• @Rogem I would say printing the array falls under accepted i/o. – Jonathan Frech Mar 6 '18 at 19:07
• @KevinCruijssen Thanks a lot; using the ease of parsing I managed to shave off another byte. – Jonathan Frech Mar 6 '18 at 19:12

# Charcoal, 5 bytes

θ‖Ｃ→↓


Try it online!

Thanks to ASCII-only for a better input format.

• I wonder if this input format is valid, since I'm afraid Charcoal can't handle input otherwise. If it isn't, I'll happily delete this answer. – Erik the Outgolfer Mar 5 '18 at 18:41
• This is valid I/o. – Pavel Mar 5 '18 at 18:42
• @Pavel I just wondered because you had said that "Your program will receive a 2 dimensional array of integers", while a string is 1-dimensional (and no, the outer [] don't exactly make it 2D). – Erik the Outgolfer Mar 5 '18 at 18:44
• @ASCII-only Charcoal really needs a better I/O method... – Neil Mar 5 '18 at 20:09
• @Neil He didn't get pinged here, but I pinged him over TNB. :) – Erik the Outgolfer Mar 5 '18 at 20:13

D,f,@,bU€{r}B]{r}
D,r,@,dbR+


Try it online!

The footer simply transforms the nested array into the format in the question. Defines a function f, which expects a matrix (nested array) as an argument.

# Julia 0.6, 55 49 bytes

~i=i:-1:1
!x=[x x[:,~end];x[~end,:] x[~end,~end]]


Try it online!

~(i) is a function to create slice from i down to 1.
So ~end gives the slice end:-1:1

!(x) is the function to do the rebuilding of the array.

# V, 12 bytes

yGæGPÎy$æ_|P  Try it online! Explanation: yG " Yank every line æG " Reverse the order of the lines P " Paste what we yanked Î " On every line: y$        "   Yank the whole line
æ_      "   Reverse the whole line
|     "   Move to the beginning of the line
P    "   Paste what we yanked
`