如何在保留索引的同时将Pandas数据帧转换为np.array？

Question

For example, I have a small set of data (from movielens)

check.csv

userId,movieId,rating,timestamp
1,31,2.5,1260759144
1,1029,3.0,1260759179
1,1061,3.0,1260759182
2,17,5.0,835355681
3,267,3.0,1298861761
3,296,4.5,1298862418
3,318,5.0,1298862121

If I do

rating = pd.read_csv('check.csv')

Y = pd.pivot_table(rating, values='rating', index=['movieId'], columns=['userId']).values

it will create 3*7 matrix. But what I want is a 3*1061 (having userId as column index and movie as row index). How to achieve that 3*1061 matrix how can I make a 1061*3 matrix S, with S[31][1]=2.5 S[1029][1]=3 etc. and all the missing entries equals zero.

Answer 1

Okay, then, I think you want this.

df = pd.read_csv('check.csv')
Y = pd.pivot_table(df, values=['rating'], index=['movieId'], columns=['userId'])

df_out = pd.DataFrame(index=np.arange(Y.index.values.max())
          ).merge(Y, left_index=True, right_index=True, how='outer'
          ).fillna(0))

Answer 2

df = pd.read_csv('check.csv')
Y = pd.pivot_table(df, values=['rating'], index=['movieId'], columns=['userId'])

            rating
 userId     1       2      3
 movieId
   31      2.5      0      0
   1029    3.0      0      0
   1061    3.0      0      0
   17       0       5.0    0
   296      0       0     4.0

remaining value will come according to csv values. More details http://pbpython.com/pandas-pivot-table-explained.html