CODEWITHSUNDEEP
pythonpandas
João Galdino
João Galdino

Reputation: 95

How compare a string and datetime in pandas?

I get the following error TypeError: unsupported operand type(s) for &: 'str' and 'DatetimeArray', I want to check if the value in column "Resumo" equals "Transfêrencia" and check a date comparison.

But with "np.select" I get a error because the type of data.

How can I do that?

>>>df
Resumo Scan    DHR_ULT_ATUALIZACAO      DT_PREVISAO
Transferência 2021-07-20 09:00:00 2021-07-16 00:00:00
Transferência 2021-07-16 08:00:00 2021-07-16 00:00:00
Transferência 2021-07-20 16:00:00 2021-07-19 00:00:00



conditions  = [ df["Prazo"]=="No prazo",
              df["Resumo"]=="Coleta",
              df["Resumo"]=="Hub",
              df["Resumo"]=="Transbordo",
              (df["Resumo"]=="Transferência" & pd.to_datetime(df["DHR_ULT_ATUALIZACAO"], format=("%Y-%m-%d %H:%M:%S")) > pd.to_datetime(df["DT_PREVISAO"], format=("%Y-%m-%d %H:%M:%S")))
             ]

choices     = [ "No prazo",
               "Coleta",
               "Hub",
               "Falta",
               "Transferência¹"
              ]

df["Desvios"] = np.select(conditions, choices, default = "Entregues")

Upvotes: 0

Views: 454

Answers (1)

Anurag Dabas
Anurag Dabas

Reputation: 24314

It would be more easier if you convert datelike column to datetime first:

df["DHR_ULT_ATUALIZACAO"]=pd.to_datetime(df["DHR_ULT_ATUALIZACAO"], format=("%Y-%m-%d %H:%M:%S"))
df["DT_PREVISAO"]=pd.to_datetime(df["DT_PREVISAO"], format=("%Y-%m-%d %H:%M:%S"))

Wrap your last condition in a bracket:

conditions  = [df["Prazo"]=="No prazo",
              df["Resumo"]=="Coleta",
              df["Resumo"]=="Hub",
              df["Resumo"]=="Transbordo",
              ((df["Resumo"]=="Transferência") & (df["DHR_ULT_ATUALIZACAO"] > df["DT_PREVISAO"]))
             ]
                #^added bracket                    ^added bracket

Upvotes: 1

Related Questions