Return the flipped version of a number

Given a one digit number, return the number "flipped". Imagine it is a seven-segment display, so 2 would be:

And when flipped horizontally it would make:

Therefore when 2 is entered, the algorithm should return 5. 6 and 9 would return 9 and 6 respectively. 8 and 0 wouldn't change, so they would just return untouched.

Numbers that don't flip, like 3, will return -1. Note that in this situation, 1 does not flip (and should return -1)

Shortest code wins!

Any language is allowed.

-
I disagree with your last point -- a 1 on a 7 segment display would simply be flipped to the other side, so 1 should nap to 1. – Jwosty Jun 17 '14 at 17:40
You should specify I/O. Do you want a function, a program or anything specific? – Dennis Jun 17 '14 at 18:03
I am confused about how to flip each digit. If 2 becomes 5, then 6 should become backwards 9, not 9. But if 6 becomes 9, then the flip is just a rotation, so 2 becomes another 2, not 5. – kernigh Jun 17 '14 at 18:41
6, 9 rotated 180 deg, 2, 5 flipped horizontally, and 1, 3 in fact are reflections of themselves across the vertical axis. – jimmy23013 Jun 17 '14 at 19:24
The translations defined in the question are not consistent at all. Why do 2 and 5 flip, but 3 doesn't? – Rynant Jun 17 '14 at 19:29

43 characters without anything fancy.

f 0=0
f 8=8
f 2=5
f 5=2
f 6=9
f 9=6
f _= -1

Got it down to 31 characters by making it a partial function.

f=([0,-1,5,-1,-1,2,9,-1,8,6]!!)
-
I laughed. Have a +1. – Seeq Jun 17 '14 at 20:03
+1 For using Haskell to do exactly what the spec. says! – recursion.ninja Jun 18 '14 at 18:48

GolfScript, 15 14

I read the spec again and found that the input must be a string.

"0.5..29.86"\?

To run:

echo -n 2 | ruby golfscript.rb a.gs

Old version(which has integer input):

[0.5..2 9.8 6]?

-
Can't believe I didn't think of that... – Dennis Jun 17 '14 at 18:04

PowerShell - 27

'0 5  29 86'.indexof(\$args)
-

Python 2.x - 28

'015..29.86'.find(`input()`)
-
Python 2.x, specifically. – Seeq Jun 17 '14 at 19:36
With Python 3, you could remove the `s and save 2 chars. – Rynant Jun 17 '14 at 19:39
Thanks TheRare and Rynant – Willem Jun 17 '14 at 19:41

JavaScript 37 36

-
use .search() and save a byte. – Ismael Miguel Jun 18 '14 at 13:03
@IsmaelMiguel Nice one, Thanks! – DarkAjax Jun 18 '14 at 13:46

CJam, 20 bytes

q25691347`"5296W"er~

Try it online.

Output

\$ for i in {0..9}; { cjam <(echo 'q25691347`"5296W"er~') <<< \$i; echo; }
0
-1
5
-1
-1
2
9
-1
8
6

How it works

q          " Read from STDIN. The leaves a string on the stack.            ";
25691347`  " Push the string '25691347'.                                   ";
"5296W"    " Push the string '5296W'.                                      ";
er         " Perform character transliteration.                            ";
~          " Evaluate the result. Examples: '2' returns 2, 'W' returns -1. ";
-

BEFUNGE 93 - 18 14 20 Bytes

I guess the commentators are right, though Befunge being a 2d language lines are kinda different. Still, in this instant, the commentators are right.

&1g01g-.
! &  #* )'

Steps:

&

Reads input as a numerical value x, and pushes it on the stack.

1g

Gets the character value c (so, like '!' = 33, or '*' = 42. An empty cell = 32) at position x, 1.

01g-.

Reads the character value of cell 0,1 (33), subtracts it from c, and outputs it as a numerical value.

-
Quite nice. Have a +1. – Seeq Jun 17 '14 at 18:47
Please correct the length: it's 20 bytes – har-wradim Jun 19 '14 at 10:05
You've actually been counting your bytes wrong. You used 19 bytes. We count newlines and spaces. But if you switch to Befunge 98, you can save one; change 1st line to: &1g'!-. – Justin Jun 19 '14 at 17:39

JavaScript (ECMAScript 6) 27

f=x=>'1060039097'[x]-(x!=6)

JavaScript (ECMAScript 5) 43

function f(x){return'1060039097'[x]-(x!=6)}

UPDATE: edc65 has suggested a much better technique. toothbrush has suggested a much better language. At this point, my primary contributions are debugging and gumption.

-
If you change it to ECMAScript 6 (supported in Firefox), you could simply do x=>[0,-1,5,-1,-1,2,9,-1,8,6][x]. – Toothbrush Jun 17 '14 at 20:17
At first, I nearly posted function(x)[0,-1,5,-1,-1,2,9,-1,8,6][x], also thanks to Firefox. I wasn't going to win anyway, so I decided I'd just stick with the highest-compatibility answer. If I start switching languages for brevity, then I'll eventually start defining my own language for every challenge I do. But I'll go ahead and mention the ECMAScript 6 version anyway, since you suggested it – Keen Jun 17 '14 at 21:02
Same concept but shorter (bye bye commas): x=>'106003907'[x]-(x!=6) – edc65 Jun 17 '14 at 22:44
@edc65 You know, I'd wanted to use a string, and I had completely blanked on the fact that I could coerce the result back to a number. A bizarre lapse. Yet I still wouldn't have come up with -(x!=6). Thank you. – Keen Jun 18 '14 at 0:08

bash 29

tr 1-9 x5xx29x86|sed s/x/-1/g

e.g.

\$ echo 0 1 2 3 4 5 6 7 8 9 | tr 1-9 x5xx29x86|sed s/x/-1/g
0 -1 5 -1 -1 2 9 -1 8 6
-
You can omit the ' around the sed expression. Also I think you can omit the g because the spec only supplies one digit at a time – Digital Trauma Jun 17 '14 at 18:55
Thanks. It's just in the example, the submission itself doesn't use '. Liking the g for longer input! – Yimin Rong Jun 17 '14 at 19:57

Kona - 29

This function returns the element x from the array 0 -1 5...

f:{0 -1 5 -1 -1 2 9 -1 8 6@x}

Examples:

> f 2
5
> f 5
2
> f 8
8
-
Is a vector by itself really allowed? – Seeq Jun 17 '14 at 19:17
@TheRare: Hmm, it does say "algorithm" so I suppose not. I'll change it and make it more like yours... – Kyle Kanos Jun 17 '14 at 19:43
Seems better, have a +1. – Seeq Jun 17 '14 at 19:53

JavaScript 36 3741

as ES6 function - 27:

f=v=>'0x'+'65b558f5ec'[v]-6
-
28: f=v=>~-[1,,6,,,3,10,,9,7][v] – nderscore Jun 17 '14 at 20:29
@nderscore You just love improving people's codes, don't you? – Seeq Jun 17 '14 at 20:56
@TheRare I'm just on a quest to find the shortest javascript code. :) If someone else has already posted a good answer, it makes more sense to me to find optimizations in it rather than post a new answer that's almost a duplicate. I'm not here to compete, just to cooperate towards achieving this goal. – nderscore Jun 17 '14 at 21:00
@nderscore I have the same mentality as long as my idea is similar enough. Anyways, nice one. – Seeq Jun 17 '14 at 21:04
@nderscore You really gave me a motivation. I'm not sure if I can make it shorter, but I'll try :) – core1024 Jun 17 '14 at 23:11

J - 28 27 bytes

You know what I like? Simplicity (28 bytes). Note that in J, _1 is negative one (-1).

f=:{&0 _1 5 _1 _1 2 9 _1 8 6

Add a little complexity and you have 27 bytes.

f=:-{&0 2 _3 4 5 3 _3 8 0 3

Example:

f 2
5
f 6
9
f 5
2
...
-

CJam - 14

Input/output version:

"0 5  29 86"q#

Stack version (assumes the number is on the stack):

[0W5WWY9W8 6]=

-
There's no Wikipedia article about CJAM, and the link just goes to a blank fiddler. Where would you find information about the language and its established releases? – Panzercrisis Jun 19 '14 at 15:38
Looks like this is it: sourceforge.net/projects/cjam – Panzercrisis Jun 19 '14 at 15:40

Sclipting, 11 characters

걄럣뉥밈결⓷方分결剩貶

Finally I have found a Windows computer with Visual Studio installed to build its interpreter. And it has defeated my GolfScript code easily.

-
Interesting, but your GolfScript answer still wins. Unless the question states otherwise, length is measured in bytes. – Dennis Sep 14 '14 at 4:17
@Dennis This was my 3rd or 4th answer on this site. I didn't know. And I think APL is an exception. – jimmy23013 Sep 14 '14 at 8:43
@Dennis And most people don't like Sclipting. :) – jimmy23013 Sep 14 '14 at 8:47
Not really an exception, we just let people choose their encoding. This answer would score 22 bytes, since that's its size using UTF-16. APL uses only 256 different characters, and there's an APL code page where once character is exactly one bytes. – Dennis Sep 14 '14 at 13:42
@Dennis Ah, you are right. – jimmy23013 Sep 14 '14 at 13:56

Perl, 27 26

Count includes the p flag

y/2569/5296/,s/[1347]/-1/

Usage:

\$ echo 7 | perl -pe y/2569/5296/,s/[1347]/-1/
-1

Wait, did Perl just beat J? :)

-
+1 I was going to post y+0-9+9a4aa70a13+;\$_=9-hex, which is the same length, but your is more original ;) – core1024 Jun 18 '14 at 20:35
@core1024 : And it just got shorter ;) – Zaid Jun 18 '14 at 20:43
It didn't: echo 1 | perl -pe y/2569/5296/,s/[347]/-1/ – core1024 Jun 18 '14 at 20:48
@core1024 : I misread the specs, should be fixed now. It's still shorter than the J solution. – Zaid Jun 18 '14 at 21:05
There really is no concise way to do this in J, as values are strictly typed and numbers can't match strings. – Seeq Jun 18 '14 at 21:10

Java - 49

long f(int a){return(0x790a300601L>>(a*4)&15)-1;}

Here, 0x790a300601 is a value stuffed with the desired outputs, with one added to make them positive. The values are stored in nibbles within the value, so a bit of shifting and masking is required to pop them out.

Java - 100 (fun option)

int f(int a){Random r=new Random(0x2000A2F47D6Fl);int b=0;for(;a>=0;a--)b=r.nextInt(11)-1;
return b;}

Not the smallest option, but a bit of fun. I found a random seed that produces the correct values when called X times (where 0 >= X <= 9).

-

Marbelous 34 bytes

}0
=6=9=2=5=0=8?0
+3-3+3-3....--

It's not the shortest solution, but it's not the longest either.

How it works

}0 spawns a marble representing the first command line input. This marble drops down the next tick, onto the =6 cell. =6 is a comparison cell, it pushes any marble with value 6 down and all others to the right. This line-up of comparison cells pushes marbles right until they equal a desired value. 0 and 8 just fall through and get printed when tehy fall off the bottom of the board, 6 and 2, and 9 and 5 first get 3 added to them, subtracted from them respectively. If a marble doesn't equal any of the desired values, it lands on the ?0 cell, which turn any marble into a 0 marble1. This marble then gets decremented and falls off the board.

1 A ?n marble technically turns any marble into a marble between 0 and n. This has the nice side effect that ?0 turns anything into 0.

-

ECMAScript 6, 24

f=x=>~-++'0n5nn29n86'[x]

Using normal JavaScript it would be 33:

-

Python - 34

f=lambda n:ord("A@F@@CJ@IG"[n])-65
-
33 f=lambda n:ord(" "[n])-3 working with space codeskulptor.org/#user34_Q7NbNvQy55_0.py – Dylan Madisetti Jun 17 '14 at 18:33
You might want to explain how and why this works – Riot Jun 17 '14 at 18:39
Using the ascii table asciitable.com the white spaces chosen are printable in python. it actually looks something like &#002;&#002;&#008;&#002;&#002;&#005;&#012;&#002;&#011;&#009; It's minus 3 for the reason that -1 leaves a null character which is no good, and minus 2 leaves a line feed which is reserved in python – Dylan Madisetti Jun 17 '14 at 18:52

Java, 58 59

int f(int i){int[]a={1,0,6,0,0,3,10,0,9,7};return a[i]-1;}

OR

int f(int i){return new int[]{1,0,6,0,0,3,10,0,9,7}[i]-1;}
-
You can make your code a bit shorter if you add 1 to each value in array and after [i] substract 1. – Barteks2x Jun 17 '14 at 18:43
@Barteks2x Good point; thanks – Ypnypn Jun 17 '14 at 19:18

MATLAB - 35

I would wrap this in a function with n as the only parameter.

f=[0,-1,5,-1,-1,2,9,-1,8,6];
f(n+1)

35 characters.

-

JavaScript 42 37

Run it on the console of your browser

-

C - 11710810677 76 bytes

a[]={1,0,6,0,0,3,10,0,9,7};main(int c,char**b){printf("%d",a[*b[1]-48]-1);}

Not the best language for golfing, but oh well...
Compile with gcc progname.c -o progname. (Ignore the warnings, or add -include stdio.h to the compile command.)

Usage: ./progname <number>

EDIT

As per @bebe's suggestion, here is an example that takes the input from STDIN instead:

C - 68 51 bytes

main(){printf("%d","106003:097"[getchar()-48]-49);}
-
using d=*b[1]-48 could be a good idea – bebe Jun 17 '14 at 20:08
@bebe Ah yes, thanks! – BenjiWiebe Jun 17 '14 at 20:46
main(){printf("%d",(int[]){1,0,6,0,0,3,10,0,9,7}[getchar()-48]-1);} sorry for bothering you this much but i find this a bit shorter. – bebe Jun 17 '14 at 20:53
You can save another character by making the array global so you don't need the cast. a[]={1,0,6,0,0,3,10,0,9,7};main(){printf("%d",a[getchar()-48]-1);} – Allbeert Jun 18 '14 at 14:18
main(){printf("%d","106003:097"[getchar()-48]-49);} 51 bytes – bebe Jun 18 '14 at 21:25

J (24, function)

(the input panel isn't playing nice. Paste the following in a python interpreter and my answer shall be revealed:)

print "f=:{&(_3+3 u:'^C^B^B^E^L^B^K\t')"
-

TI-BASIC, 24 22

-1+int(11fPart(11^Ans.0954191904

This encodes the possible outputs in a lookup table stored as a base-11 floating point number; the (N+1)th base-11 digit after the decimal point is extracted from the table to get the value of the inverted digit. In base 11 the number is .106003A097, and the digits of this less one are exactly 0,-1,5,-1,-1,2,9,-1,8,6.

edc65's trick of subtracting one in the case of 6 leads to this 24-byte solution, where ⑩^( is a single one-byte token:

-(Ans≠6)+int(10fPart(.1060039097⑩^(Ans

The string approach is 29 bytes:

-(Ans≠6)+expr(inString("1060039097",Ans+1,1

The array approach (which Ypnpyn also took) is 30 bytes, assuming the number is stored in X:

{1,0,6,0,0,3,10,0,9,7}-1:Ans(X+1

24 -> 22: Removed two extra digits of precision in the magic constant.

-
This can almost certainly be golfed to 19, and possibly to 18; however, I need to search for the correct numbers. – lirtosiast Jun 16 '15 at 5:47

Python - 92

f=lambda x:x if x in(0,8)else(5if x==2 else(2if x==5 else(6if x==9 else(9if x==6 else -1))))

Usage:

>>> f(2)
5
>>> f(3)
-1
-
def f(x): will make it a byte shorter. Also "[a,b][x==5]" is shorter then and equivalent to "a if x==5 else b" – Pinna_be Jun 17 '14 at 21:42
Also, a small change would be to put the if x==9 last and say x>8 instead, saving a character. – Pinna_be Jun 17 '14 at 21:49
@Pinna_be: You’re forgetting the return. – Ryan O'Hara Jun 18 '14 at 5:09

Python - 73 bytes

r=[2,5,2,9,6,9,8,8,0,0]
f=lambda x:-1 if x not in r else r[r.index(x)+1]

Or, for one digit numbers only :

r=[0,-1,5,-1,-1,2,9,-1,8,6]
f=lambda x:r[x]
-
why not altogether f=(lambda x:[-1,-1,5,-1,-1,2,9,-1,8,6][x])(x) – Dylan Madisetti Jun 17 '14 at 18:24
43 def f(x):print[-1,-1,5,-1,-1,2,9,-1,8,6][x] – Dylan Madisetti Jun 17 '14 at 18:56

PowerShell 32

(0,-1,5,-1,-1,2,9,-1,8,6)[\$args]
-

F# - 37 bytes

let f n=[0;-1;5;-1;-1;2;9;-1;8;6].[n]

Usage:

> f 2
val it : int = 5
> f 9
val it : int = 6
> f 4
val it : int = -1
-
Fails for 0 and 8 – Dancrumb Jun 17 '14 at 22:30
I missed that in the rules, thanks! – Jwosty Jun 17 '14 at 22:31

Java, 75

int f(int i){return i>8?6:i>7?8:i>6?-1:i>5?9:i>4?2:i>2?-1:i>1?5:i>0?-1:0;}
-
i>8:-1 will return -1 for input 9 when it should return 6, won't it? – corsiKa Jun 17 '14 at 21:52
@corsiKa Indeed. – Ypnypn Jun 17 '14 at 23:47