无重复和还原的两个数据框组合，高效蟒蛇

Question

我有两个具有数千行的数据框，我需要将两个数据框组合成一个数据框，而无需重复和还原。例如：

数据框1

drug1
drug2
drug3

数据框2

disease1
disease2
disease3

因此，输出数据帧将为：

输出数据帧

drug1 disease1
drug1 disease2
drug1 disease3
drug2 disease1
drug2 disease2
drug2 disease3 
drug3 disease1
drug3 disease2
drug3 disease3

我不想要包含以下内容的输出组合：

disease1 drug1
drug1 drug1
disease1 disease1 

我实际上使用pd.merge进行了尝试，但是它返回重复和还原，并且还花费了很长时间，因为我在数据框1和2中有成千上万个

请帮忙吗？

Answer 1

仅在pandas中使用的一种方法是创建一个MultiIndex from product，然后将其转换为数据框：

>>> df1
       0
0  drug1
1  drug2
2  drug3
>>> df2
          0
0  disease1
1  disease2
2  disease3

df3 = (pd.MultiIndex.from_product([df1[0],df2[0]])
       .to_frame()
       .reset_index(drop=True))

>>> df3
       0         1
0  drug1  disease1
1  drug1  disease2
2  drug1  disease3
3  drug2  disease1
4  drug2  disease2
5  drug2  disease3
6  drug3  disease1
7  drug3  disease2
8  drug3  disease3

Answer 2

设置

public class MyRecyclerViewActivity extends AppCompatActivity {
RecyclerView recyclerView;

@Override
protected void onCreate(@Nullable Bundle savedInstanceState) {
    super.onCreate(savedInstanceState);
    setContentView(R.layout.my_recycler_view);

    DatabaseReference databaseReference = FirebaseDatabase.getInstance().getReference();


    databaseReference.addChildEventListener(new ChildEventListener() {
        @Override // indivivdual items at the db ref
        public void onChildAdded(DataSnapshot dataSnapshot, String s) {

            String fileName = dataSnapshot.getKey(); //returns file name
            String url = dataSnapshot.getValue(String.class); //returns url for file name

            ((MyAdapter) recyclerView.getAdapter()).update(fileName,url);


        }

        @Override
        public void onChildChanged(DataSnapshot dataSnapshot, String s) {

        }

        @Override
        public void onChildRemoved(DataSnapshot dataSnapshot) {

        }

        @Override
        public void onChildMoved(DataSnapshot dataSnapshot, String s) {

        }

        @Override
        public void onCancelled(DatabaseError databaseError) {

        }
    });

    recyclerView = findViewById(R.id.recyclerView);

    recyclerView.setLayoutManager(new LinearLayoutManager(MyRecyclerViewActivity.this));

    MyAdapter myAdapter = new MyAdapter(recyclerView,MyRecyclerViewActivity.this, new ArrayList<String>(), new ArrayList<String>());
    recyclerView.setAdapter(myAdapter);

}
}

---

df1 = pd.DataFrame(dict(col1=[f"drug{i}" for i in range(1, 4)])) df2 = pd.DataFrame(dict(col2=[f"disease{i}" for i in range(1, 4)])) 在分配的列上

merge

---

理解力

df1.assign(A=1).merge(df2.assign(A=1)).drop('A', 1)

    col1      col2
0  drug1  disease1
1  drug1  disease2
2  drug1  disease3
3  drug2  disease1
4  drug2  disease2
5  drug2  disease3
6  drug3  disease1
7  drug3  disease2
8  drug3  disease3

---

pd.DataFrame([ (i, j) for i in df1.col1 for j in df2.col2 ], columns=['col1', 'col2'])

归纳为任意两个数据框的叉积

pandas.concat

Answer 3

尝试以下解决方案：

i = df1.index.repeat(len(df2))
j = np.tile(df2.index, len(df1))

pd.concat([
    df1.loc[i].reset_index(drop=True),
    df2.loc[j].reset_index(drop=True)
], sort=True, axis=1)