Alphabet triangle strikes again

A
BCD
EFGHI
JKLMNOP
QRSTUVWXY
ZABCDEFGHIJ
KLMNOPQRSTUVW
XYZABCDEFGHIJKL
MNOPQRSTUVWXYZABC
DEFGHIJKLMNOPQRSTUV
WXYZABCDEFGHIJKLMNOPQ
RSTUVWXYZABCDEFGHIJKLMN
OPQRSTUVWXYZABCDEFGHIJKLM
NOPQRSTUVWXYZABCDEFGHIJKLMN
OPQRSTUVWXYZABCDEFGHIJKLMNOPQ
RSTUVWXYZABCDEFGHIJKLMNOPQRSTUV
WXYZABCDEFGHIJKLMNOPQRSTUVWXYZABC
DEFGHIJKLMNOPQRSTUVWXYZABCDEFGHIJKL
MNOPQRSTUVWXYZABCDEFGHIJKLMNOPQRSTUVW
XYZABCDEFGHIJKLMNOPQRSTUVWXYZABCDEFGHIJ
KLMNOPQRSTUVWXYZABCDEFGHIJKLMNOPQRSTUVWXY
ZABCDEFGHIJKLMNOPQRSTUVWXYZABCDEFGHIJKLMNOP
QRSTUVWXYZABCDEFGHIJKLMNOPQRSTUVWXYZABCDEFGHI
JKLMNOPQRSTUVWXYZABCDEFGHIJKLMNOPQRSTUVWXYZABCD
EFGHIJKLMNOPQRSTUVWXYZABCDEFGHIJKLMNOPQRSTUVWXYZA
BCDEFGHIJKLMNOPQRSTUVWXYZABCDEFGHIJKLMNOPQRSTUVWXYZ


Specs

• You may do it in all-lowercase instead of all-uppercase.
• Trailing newlines at the end of the triangle is allowed.
• Trailing spaces after each line is allowed.
• You must print to STDOUT instead of outputting an array of strings.

Scoring

This is . Program with lowest byte-count wins.

• What do you mean by "strikes again"? Was there another challenge you made like this? Aug 21, 2016 at 13:46
• Aug 21, 2016 at 13:58
• Seems fairly trivial do we really need (another) alphabet challenge? Aug 21, 2016 at 21:44
• It is a good challenge, but I think we have outstripped saturation of these alphabet challenges, nothing personal. Aug 21, 2016 at 21:45
• Actually looking for an alphabet challenge that the letter at a position cannot be calculated by simple expressions from its coordinates involving the mod function. May make one myself if I have time. Jan 31, 2018 at 20:17

Vim, 29 bytes

:h<_↵↵↵y$ZZ26P0qqa↵♥βjllq25@q  Where ↵ represents the Return key, ♥ the Escape key, and β the Backspace key. • How do you always beat me to the vim answer on these? Aargh +1 anyway, I can't not upvote vim! :) Aug 21, 2016 at 12:41 • I still think you should use ␛ instead of ♥. And ␈ instead of β. That's what these Unicode chars were made for. utf8icons.com/subsets/control-pictures Aug 23, 2016 at 20:36 Python 2, 65 bytes i=1 a=bytearray(range(65,91))*26 while a:print a[:i];a=a[i:];i+=2  • I've changed the header to Python 2 because the code would not work in Python 3. Aug 21, 2016 at 11:49 Jelly, 10 bytes 26RḤ’RØAṁY  Try it online! How it works 26RḤ’RØAṁY Main link. No Arguments. 26 Set the return value to 26. R Range; yield [1, 2, ..., 25, 26]. Ḥ Unhalve; yield [2, 4, ..., 50, 52]. ’ Decrement; yield [1, 3, ..., 49, 51]. R Range; yield [[1], [1, 2, 3], ..., [1, ..., 49], [1, ..., 51]]. ØA Yield the uppercase alphabet. ṁ Mold the alphabet as the array of ranges. This repeats the uppercase letters over an over again, until all integers in the range arrays have been replaced with letters. Y Join, separating by linefeeds.  • Was "Double" called "Unhalve" back then? Also, ṁ is great!! [waiting to say congrats for 100k rep] Sep 28, 2016 at 11:14 • It's just a mnemonic. H is halve and Ḥ is its inverse (unhalve). Sep 28, 2016 at 12:15 • I just think of /2 or *2, so it's "Halve" or "Double". That's why I was confused. Sep 28, 2016 at 12:17 • Also 10 bytes: 27Ḷ²IRØAṁY Apr 30, 2017 at 4:32 • Also 10 bytes: 51Rm2RØAṁY Apr 30, 2017 at 5:01 VBA Excel (80 bytes, 1742 bytes) Excel, 1742 bytes Inspired by the ugoren's creative answer, I managed to find an Excel formula to create the pattern as shown in the OP. =MID(REPT("ABCDEFGHIJKLMNOPQRSTUVWXYZ",26),(ROW()-1)^2+1,2*ROW()-1)  Paste this formula in cell A1, then drag all over range A1:A26. The length of the formula is 67 bytes but you have to replicate it 26 times, so it's equal to 67*26=1742 bytes. Here is the output: Excel VBA, 80 bytes Now it's possible we integrate Excel with VBA to automate the process and to save many bytes since VBA is built into most Microsoft Office applications, including Excel. Write and run the following code in the Immediate Window (use combination keys CTRL+G to display it in Visual Basic Editor): [A1:A26]="=MID(REPT(""ABCDEFGHIJKLMNOPQRSTUVWXYZ"",26),(ROW()-1)^2+1,2*ROW()-1)"  The program works by printing the Excel formula above to the range A1:A26. Unfortunately, both Excel and VBA have no built-in alphabet. • The column names look like a built-in alphabet to me. Use the first 26 column names. Aug 23, 2016 at 20:32 • @mbomb007 So what? I think it would be harder to use them instead of strings. Sep 28, 2016 at 11:18 • @EriktheGolfer So? My point was that there is a builtin alphabet. Sep 28, 2016 at 13:26 • @mbomb007 You said "Use the first 26 column names", which I perceived as "Use the first 26 column names instead of what you currently use", that's why I replied. Sep 28, 2016 at 13:29 • @EriktheGolfer It's a suggestion. Idk how many bytes it'd be. Sep 28, 2016 at 13:30 Haskell, 67 bytes _#53=[] s#i=take i s:drop i s#(i+2) mapM putStrLn$cycle['A'..'Z']#1


A simple recursion over the length i of the line. In each step the next i chars are taken from an infinite repetition of the alphabet.

Mathematica, 90 bytes

StringRiffle[Flatten[Alphabet[]&~Array~26]~InternalPartitionRagged~Range[1,51,2],"
",""]&


Anonymous function. Takes no input and returns a string as output. Golfing suggestions welcome. An example of what InternalPartitionRagged does:

In[1]:= InternalPartitionRagged[{2, 3, 5, 7, 11, 13}, {2, 3, 1}]

Out[1]= {{2, 3}, {5, 7, 11}, {13}}

• Mathematica has a built-in for alphabet triangles? Aug 21, 2016 at 18:35

C, 79 bytes

main(i,j){for(i=0,j=1;i<676;i++){putchar(i%26+65);if(j*j==i+1){puts("");j++;}}}


My first answer in C \o/

Golfing suggestions are more than welcome.

• 62: i;main(j){while(i<676)printf("\n%c"+(j*j^i++||!j++),i%26+65);}
– xsot
Aug 21, 2016 at 12:37
• @xsot Thanks but I'm afraid that leading newlines are not allowed. Aug 21, 2016 at 12:47
• But there's no leading newline?
– xsot
Aug 21, 2016 at 12:50
• 60: i;main(j){for(;j<27;j*j^++i||puts("",j++))putchar(i%26+65);}
– xsot
Aug 21, 2016 at 13:00
• @immibis I guess I should post it then.
– xsot
Aug 22, 2016 at 23:57

Brachylog, 37 bytes

26~l<:1aLc~j[@A:I],L~@nw
=:2%1,.#@l?,


Try it online!

Explanation

• Main predicate:

26~l         Let there be a list of 26 elements
<            This list is an ascending list of integers
:1aL         Apply predicate 1 to that list ; the resulting list of strings is L
c            Concatenate the list of strings into one big string
~j[@A:I],    That big string is the result of juxataposing the alphabet I times to itself
L~@n         Create a string which when splitted on line breaks results in L
w            Write that string to STDOUT

• Predicate 1: used to generate variable strings of odd lengths.

=            Assign a value to the Input
:2%1,        That value must be odd
.#@l?,       Output is a string of length Input

• Finally you do an ascii-art challenge Aug 21, 2016 at 12:30
• @LeakyNun Would classify that as string manipulation more than ASCII art imo Aug 21, 2016 at 12:33
• Should I add that to the tag? Aug 21, 2016 at 12:43
• Nice use of the fact that the last line is the only line which ends in Z which is because 26 is square-free. Aug 21, 2016 at 12:44

Perl, 4241 39 bytes

perl -E '@b=(A..Z)x26;say splice@b,0,$#a+=2for@b'  Just the code: @b=(A..Z)x26;say splice@b,0,$#a+=2for@b


An obvious shorter version unfortunately triggers an internal perl problem (Use of freed value in iteration):

say splice@b,0,$#a+=2for@b=(A..Z)x26  Pyramid Scheme, 1492 bytes  ^ ^ / \ / \ /set\ /do \ ^-----^ ^-----^ /a\ ^- /!\ /[\ --- ^- ^--- ^---^ ^- /=\ ^- / \ ^- ^---^ -^ /out\ ^- ^- /a\ ^- -----^ ^- / \ ---/[\ / \ -^ /51 \ ^---^ /chr\ ^- ----- ^- / \ -----^ -^ ^- /do \ / \ ^- ^- ^-----^ /10 \ -^ /[\ -^ ^- ----- ^- ^---^ -^ -^ -^ / \ -^ /c\ ^- ^- /set\ /[\--- -^ ^- ^-----^---^ ^- -^ /a\ /+\ -^ -^ ^- --- ^---^ -^ -^ -^ /a\ /2\ / \ -^ -^ --- ---/set\ -^ -^ ^-----^ ^- -^ /c\ /a\-^ -^ --- --- -^ -^ /[\ -^ ^---^ -^ ^- -^ /]\ / \ -^ ^---^ /set\ -^ ^- -^ ^-----^ -^ / \ ^- /c\ /-\ -^ /set\ ^- --- ^---^ -^ ^-----^- /c\ /1\ -^ /b\ / \ --- --- /]\ --- / -1\ ^---^ ----- ^- / \ / \ /out\ /set\-----^ ^-----^ -^ /b\ /?\ -^ --- ^---^ / \ /!\ -^ /chr\ ^--- /+\-----^ /=\ ^---^ /+\ ^---^ /b\ /1\ ^---^ /b\ / \--- ---/ \ /b\ ---/25 \ /65 \--- ----- -----  Try it online! Why? Why not? Pyth, 12 bytes jtc*G26*Rd26  Try it online! JavaScript (ES6), 77 82 88 EcmaScript 6 required just to save 1 byte using a template string literal for newline. for(i=r=l=o='';l+52;r++||(r=l-=2,o+= ))o+=(i++%26+10).toString(36);alert(o)  Less golfed for(i = r = l = o = ''; l + 52; r++ || (r = l -= 2, o += \n)) o += (i++ % 26 + 10).toString(36); alert(o);  Test for(i=r=l=o='';l+52;r++||(r=l-=2,o+= ))o+=(i++%26+10).toString(36);alert(o) • \o/ an ES6 answer I can run in my browser Aug 21, 2016 at 19:53 • I stole your .toString(36) logic...now you have to beat 80 bytes!! Aug 22, 2016 at 12:19 • Ahh I concede, I can't think of any way to beat this with my library. Thanks for the challenge! Aug 22, 2016 at 15:57 • 69 – l4m2 Mar 19 at 4:17 Husk, 10 bytes Cİ1ṠṁK…"AZ  Try it online! Explanation Cİ1ṠṁK…"AZ …"AZ Get the alphabet ṠṁK Replace each letter with the whole alphabet C Cut the resulting string into lines with lengths İ1 equal to the list of odd numbers  JavaScript, 129 bytes z=1,i=0,g=a=>{b=0,c="";while(a+a-1>b){c+='ABCDEFGHIJKLMNOPQRSTUVWXYZ'.split[i>26?i=0:i++];b++}console.log(c)};while(z<26)g(z++)  • Using the spread operator could save you 3 bytes: [...'ABCDEFGHIJKLMNOPQRSTUVWXYZ'] instead of 'ABCDEFGHIJKLMNOPQRSTUVWXYZ'.split. Aug 21, 2016 at 18:46 Pyke, 14 bytes G26*WDoh< Ko>D  Try it here! Go, 133 bytes package main import S "strings" func main(){s:=S.Repeat("ABCDEFGHIJKLMNOPQRSTUVXXYZ",26) for i:=1;s!="";i+=2{println(s[:i]);s=s[i:]}}  MATLAB, 112 109 95 79 77 bytes This will also work with Octave, you can try online here. a=['' 65:90 65:90]';j=0;for i=1:2:52;b=circshift(a,j);j=j-i;disp(b(1:i)');end  So after some major changes, I've saved a further 14 32 bytes. This one is getting to be much more like the length I would have expected from MATLAB. I've left the old version below as it is substantially different. a=['' 65:90 65:90]'; %Create 2 copies of the upper case alphabet j=0; %Initialise cumulative sum for i=1:2:52; %For each line length b=circshift(a,j); %Rotate the alphabet string by the cumulative offset j=j-i; %Update cumulative negative sum of offsets. disp( b(1:i)' %Extract first 'line length' characters from rotated alphabet. ); %Display next line (disp adds newline automatically) end  Original version: a=['' repmat(65:90,1,26)];c=cumsum(1:2:51);disp(cell2mat(arrayfun(@(s,f)[a(s:f) 10],[1,c(1:25)+1],c,'Un',0)))  Wow that one ended up being longer than I thought it would. I'll see if I can't knock a few bytes off it. An ungolfed version to explain: a=['' repmat(65:90,1,26)]; %Create 26 copies of the upper case alphabet c=cumsum(1:2:51); %Calculate the end index of each row in the above array, by cumulatively summing the length of each line disp( cell2mat( arrayfun(@(s,f) [a(s:f) 10], %Extract the part of the alphabet and insert a new line. [1,c(1:25)+1],c, %start index is the previous end of line plus 1. End index is as calculated by cumsum. 'Un',0 %The returned values are not all the same length ) %Do this for each line ) %Convert back to a matrix now new lines inserted ) %And display it  Acknowledgements • 3 bytes saved - thanks @LuisMendo XPath 3.0 (and XQuery 3.0), 84 bytes codepoints-to-string((0 to 25)!(subsequence(((1 to 26)!(65 to 90)),.*.+1,2*.+1),10))  Explanation: (1 to 26)!(65 to 90) is the alphabet 26 times (0 to 25)!(subsequence(XX, start, len),10) takes 26 subsequences of this, each followed by newline subsequence(X, .*.+1, 2*.+1) takes successive subsequences with start position and length: (1, 1), (2, 3), (5, 5), (10, 9) etc. codepoints-to-string() turns Unicode codepoints into characters • Bravo. I thought I knew what XQuery was. Turns out I had no idea. Aug 22, 2016 at 4:25 • I doubt that many of the posts here tell you much about the language they are written in. Aug 22, 2016 at 21:24 Ruby, 46 bytes 26.times{|i|puts ([*?A..?Z]*26)[i*i,i*2+1]*""}  See it on ideone: http://ideone.com/3hGLB0 05AB1E (alternate) 15 bytes A2×52µ¼D¾£,¾¼FÀ  Try it online! Explanation: A2× # push a string containing a-za-z 52µ # Loop the rest of the program until counter = 52 ¼ # increment counter (it's 0 initially) D # Duplicate the alpha string on the stack ¾£ # Replace alpha with alpha[0..counter] , # Pop the substring and print it ¾¼FÀ # rotate the alpha string left counter++ times.  R, 120115 111 bytes v=c();for(i in 1:26){v=c(v,c(rep(LETTERS,26)[(sum((b=seq(1,51,2))[1:i-1])+1):sum(b[1:i])],"\n"))};cat(v,sep="")  Ungolfed : a=rep(LETTERS,26) b=seq(1,51,2) v=vector() for(i in 1:26) { v=c(v,c(a[(sum(b[1:i-1])+1):sum(b[1:i])],"\n")) } cat(v,sep="")  Basically, b is the vector of the odd numbers between 1 and 51, thus giving the length of each line. Obviously, the sum function sums the numbers of this vector, and gives the starting and ending indexes. -5 bytes thanks to @plannapus ! -4 bytes thanks to @plannapus ! • arf, sorry I didn't see that earlier, but since you're only using a once you don't actually need to define it, meaning you can shave a couple more bytes: b=seq(1,51,2);v=c();for(i in 1:26){v=c(v,c(rep(LETTERS,26)[(sum(b[1:i-1])+1):sum(b[1:i])],"\n"))};cat(v,sep="") works. Aug 23, 2016 at 13:52 • @plannapus Silly me ! Thanks again ! I also integrated the b=seq part in the main body, so it's even less readable ! Aug 23, 2016 at 13:58 R, 81 73 65 63 bytes A simple for loop approach. Repeat the alphabet 26 times and loop through a sliding index range that is calculated using (i^2-2*i+2):i^2. for(i in 1:26)cat(rep(LETTERS,26)[(i^2-2*i+2):i^2],"\n",sep="")  Batch, 123 bytes @set s= @for /l %%i in (1,2,51)do @call set s=%%s%%ABCDEFGHIJKLMNOPQRSTUVWXYZ&call echo %%s:~0,%%i%%&call set s=%%s:~%%i%%  05AB1E, 18 17 bytes 26FA26×N>n£NnF¦},  Explanation 26F # for N in range(0, 26) do A26× # the alphabet repeated 26 times N>n£ # take (N+1)^2 letters NnF¦} # throw away the first N^2 letters , # print with newline  Try it online! Rexx, 74 72 bytes i=1;m=1;do 26;say substr(copies(xrange('A','Z'),26),i,m);i=i+m;m=m+2;end  Ungolfed: i=1 m=1 do 26 say substr(copies(xrange('A','Z'),26),i,m) i=i+m m=m+2 end  TSQL, 129 bytes USE MASTER in the beginning of the script is to ensure that the query is run in the master database which is default for many users(not counting bytes for that). Golfed: USE MASTER SELECT SUBSTRING(REPLICATE('ABCDEFGHIJKLMNOPQRSTUVWXYZ',26),number*number+1,number*2+1)FROM spt_values WHERE number<26and'P'=type  Ungolfed: USE MASTER SELECT SUBSTRING(REPLICATE('ABCDEFGHIJKLMNOPQRSTUVWXYZ',26),number*number+1,number*2+1) FROM spt_values WHERE number<26and'P'=type  Fiddle Fiddle for older version using xml path PowerShell, 68 bytes $a=-join(65..90|%{[char]$_})*26;26..1|%{$a=$a.Insert($_*$_,"n")};$a


The section before the first semicolon produces a string containing 26 copies of the uppercase alphabet. The next section injects linebreaks at the index of each square number (working backward so I don't have to account for the shifting). Finally, the \$a at the end just shoves that string variable onto PowerShell's equivalent of STDOUT.

Dyalog APL, 18 bytes

↑2{⍺↓⍵⍴⎕a}/×⍨0,⍳26

• 18 bytes or 18 characters? To encode these 18 characters as 18 bytes, you need a custom character encoding; and we could compress any of the solutions if we choose our character encoding carefully enough. Aug 21, 2016 at 23:06
• Aug 22, 2016 at 0:09
• Will you explain this, or shall I?
Aug 23, 2016 at 11:26

C, 60 bytes

i;main(j){for(;j<27;j*j^++i||puts("",j++))putchar(i%26+65);}

• puts only takes one argument. (Some undefined behaviour is permitted in codegolf normally but this is a bit too far outside the usual lanes)
– M.M
Aug 23, 2016 at 1:13
• @M.M Undefined behaviour is exploited all the time. The rule is that a submission is valid as long as it works in some compiler, otherwise we would have to explicitly rule out a long list of exceptions. This code works in gcc so it's a valid submission.
– xsot
Aug 23, 2016 at 1:38

C++, 111 bytes

void a(){int c=65,i,j;for(i=0;i<26;i++){for(j=0;j<=2*i;j++){std::cout<<(char)c;c++;if(c==91)c=65;}std::cout<<'\n';}}


First try at one of these. Uses an int "c" to record which letter it needs to print at any given time. Once "c" passes 90 ('Z') it gets reset to 65 ('A'). Prints the pyramid using for loops.

• Nice answer! You could do if(c<92)c=65 to take one byte off, and you might also be able to do int a() instead of void a(), but I'm not positive if that works without the return. Other than that, I think you need to include #include <iostream> in your byte count. Aug 23, 2016 at 6:05
• I believe you meant if(c>90)c=65, but thanks for the suggestion, it's a good idea. Also, I guess I'll include it, thanks. Aug 26, 2016 at 8:02