**DF1 **
| Unique IDs | Task | Date 1 | Date 2 |
|---|---|---|---|
| 123 | Track | 28/05/2018 | 28/05/2018 |
| 455 | Expire | 07/02/2018 | |
| 986 | Sell | 05/01/2020 | 28/05/2018 |
**DF2 **
| Unique IDs |
|---|
| 123 |
| 455 |
| 986 |
What I’d like to do is, look up the unique ID from DF2 and get the Date 1 that corresponds to the "Track" Task. If the Track task is empty then take the Date 2.
Desired Output
| Unique IDs | Date |
|---|---|
| 123 | 28/05/2018 |
| 455 | 07/02/2018 |
| 986 | 05/01/2020 |
Ideally I would like an if statement
>Solution :
You can do it with a mask, to know if Date1 is nan or not and then set date column of df2 either with date1 values or date2 values.
df1 = pd.DataFrame({"id": [123, 455, 986],
"task": ["Track", "Expire", "Sell"],
"date1": [pd.Timestamp("28/05/2018"), np.nan, pd.Timestamp("05/01/2020")],
"date2": [pd.Timestamp("28/05/2018"), pd.Timestamp("07/02/2018"), pd.Timestamp("28/05/2018")]})
df2 = pd.DataFrame({"id": [123, 455, 986]})
mask = df1.loc[:, "date1"].notna()
df2.loc[mask, "date"] = df1.loc[mask, "date1"]
df2.loc[~mask, "date"] = df1.loc[~mask, "date2"]
print(df2)
id date
0 123 2018-05-28
1 455 2018-07-02
2 986 2020-05-01