Move columns to the right in Pandas-CodePudding

I have dataset:

Stop_ID	On_Demand	Tarif	Heading
1	T100	Station
2	DYes	T101	Gym
3	T101	River

I want to have:

Stop_ID	On_Demand	Tarif	Heading
1	DNo	T100	Station
2	DYes	T101	Gym
3	DNo	T101	River

I tried:

for index, row in sm.iterrows():
    if not row["On_Demand"].startswith("D"):
        sm.loc[index, "Tarif":] = row["On_Demand":]
        sm.loc[index, "On_Demand"] = "DNo"

but it doesnt bring any result. Any idea how can I achieve it?

CodePudding user response：

Use DataFrame.shift with axis=1 only filtered rows by inverted mask by ~ and then set column On_Demand by same mask:

m = df["On_Demand"].str.startswith("D")

df.loc[~m, "On_Demand":] = df.loc[~m, "On_Demand":].shift(axis=1)
df.loc[~m, "On_Demand"] = "DNo"

print (df)
   Stop_ID On_Demand Tarif  Heading
0        1       DNo  T100  Station
1        2      DYes  T101      Gym
2        3       DNo  T101    River