This dataset contains a million of Rubik's cube status along with the best next move for each status.
This dataset is generated by this notebook, check it out for more information:
https://www.kaggle.com/code/marcellosemboli/one-million-of-rubik-cube-moves

The cube

Notations below are taken from World Cube Association.

Faces of the cube

A solved cube has 6 faces and 6 colors:

• U - upper, White
• L - left, Orange
• F - front, Green
• R - right, Red
• B - back, Blue
• D - down, Yellow

Faces are coded as follows:

• U - 0
• L - 1
• F - 2
• R - 3
• B - 4
• D - 5

Colors

Colors are coded as follow:

• 0 White
• 1 Orange
• 2 Green
• 3 Red
• 4 Blue
• 5 Yellow

Each face has 9 tiles (from 0 to 8) for a total of 6 x 9 = 54 tiles.

Names of tiles are U0,..,U8,L0,..,L8,..,D8

Status of cube

A status of cube is a tuple of 54 colors, 9 tiles for each of 6 faces.

According to the python rubik-cube module a solved cube status is:

WWWWWWWWWOOOGGGRRRBBBOOOGGGRRRBBBOOOGGGRRRBBBYYYYYYYYY

In numbers:

000000000111222333444111222333444111222333444555555555

The status of the same cube after L move is:

400400400111022333445111022333445111022333445255255255

See below for other examples.

Moves

Rubik's cube has 12 possible moves, 6 clockwise and 6 counterclockwise, coded as follows:

Clockwise

• U - 0
• L - 1
• F - 2
• R - 3
• B - 4
• D - 5

Counterclockwise

• U' - 6
• L' - 7
• F' - 8
• R' - 9
• B' - 10
• D' - 11

Double moves as U2, L2 etc. are not used.

Scramble

A scramble is a sequence of moves, where subsequent moves are'nt opposite. Sequences of three same moves are excluded as well.

Since the God's number is 20, I will use scrambles of maximum 20 moves and the opposite of last move is the best move for a solve of that scramble.

The dataset

The dataset is a csv file were each line is a status of the cube, and the best move for that status.
Columns from 0 (the first) to 53 are named after tiles position in status of the cube, and contains the color number.
Last 12 columns are named after moves and are all 0 except the best move is 1.

U0,U1,U2,U3,U4,U5,U6,U7,U8,L0,L1,L2,F0,F1,F2,R0,R1,R2,B0,B1,B2,L3,L4,L5,F3,F4,F5,R3,R4,R5,B3,B4,B5,L6,L7,L8,F6,F7,F8,R6,R7,R8,B6,B7,B8,D0,D1,D2,D3,D4,D5,D6,D7,D8,U,L,R,F,B,D,Ui,Li,Ri,Fi,Bi,Di

For a total of 54 + 12 = 66 columns.

Each row is 132 bytes, one million rows are 132M bytes = 126 Mb.