Merge dataframes with diferent number of rows in Python

Question

I initially have a dataframe, df, that looks like: (these are country temperature data)

|date      | Brasil | Colombia | Peru | Mexico|
------------------------------------------------
|2020-09-25|1       |4         |3      |2     |
|2020-09-26|3       |4         |3      |1     |
|2020-09-27|7       |5         |4      |3     |
|2020-09-28|0       |2         |4      |6     |
|2020-09-29|2       |4         |3      |4     |
|2020-09-30|1       |2         |3      |4     |

And I have to add the dates missing till today, these dates come from dataframes for each country, which are in a directory of my computer and look like this, for Colombia as an example:

|datetime    |temp  |temp_max |
------------------------------
|2020-10-01  |5     |23       |   
|2020-10-02  |6     |22       |
|2020-10-01  |9     |23       |
|2020-10-03  |4     |25       |
|2020-10-04  |2     |4        |
|2020-10-05  |4     |12       |
|2020-10-06  |1     |11       |

the name of these dataframe on the directory is like this: Colombia_20201006. I made the next code to read the files and complete the df dataframe.

columnas = numpy.array(df.columns)
columnas = numpy.delete(columnas,0)


path = "D:/Users/user/Documentos/"

lista_archivos = os.listdir(path)

for j in columnas: #Iternado por ciudades
    paises_archivos = []
    for i in lista_archivos:
        paises_activos.append(path+i) if i.startswith(f"{j}") else print("", end="")
    print(f"los archivos para el pais {j} son: {paises_archivos}")

    for g in paises_archivos:
        paises_concat = pd.read_csv(f"{g}")
        paises_concat = paises_concat[["datetime","temp"]]
        df_out = df.merge(paises_concat,how="outer", left_on="date", right_on="datetime")
        df_out = df_out.fillna("")
        df_out.loc[(df_out.date == "") & (df_out.datetime != ""), "date"] = df_out["datetime"] 
        df_out.loc[(df_out[f"{j}"] == "") & (df_out.temp != ""), f"{j}"] = df_out["temp"] 
        df_out = df_out.drop(["datetime", "temp"], axis=1)

The problem is that not necessarily all files are the same and there may be some that do not have the same date, for example there may (this is a possibility, not a certainty) be some that, for example, do not have data between 2020-10-01 and 2022-01-31, and their dataframe looks like this (Let's assume this is the case for Peru):

|datetime    |temp  |temp_max |
------------------------------
|2022-02-01  |5     |13       |   
|2022-02-02  |8     |12       |
|2022-02-03  |9     |13       |
|2022-02-04  |1     |15       |
|2022-02-05  |2     |14       |
|2022-02-06  |5     |14       |
|2022-02-07  |1     |15       |

And what I need is that it fills the missing spaces until 2022-02-01 with NA and looks something more or less like this:

|date      | Brasil | Colombia | Peru | Mexico|
------------------------------------------------
|2020-09-25|1       |4         |3      |2     |
|2020-09-26|3       |4         |3      |1     |
|2020-09-27|7       |5         |4      |3     |
|2020-09-28|0       |2         |4      |6     |
|2020-09-29|2       |4         |3      |4     |
|2020-09-30|1       |2         |3      |4     |
|2020-10-01|1       |2         |NA     |4     |
|2020-10-02|1       |2         |NA     |4     |
|2020-10-03|1       |2         |NA     |4     |
|2020-10-04|1       |2         |NA     |4     |
|...       |...     |...       |...    |...   |
|2022-02-01|1       |5         |5      |6     |
|2022-02-02|1       |6         |8      |4     |
|2022-02-03|1       |9         |9      |8     |
|2022-02-04|1       |4         |1      |2     |
|2022-02-05|1       |2         |2      |1     |
|2022-02-06|1       |4         |5      |4     |
|2022-02-07|1       |1         |1      |3     |

But right now what the code is doing is leaving the dataframe as follows:

|date      | Brasil | Colombia | Peru | Mexico|
------------------------------------------------
|2020-09-25|1       |4         |3      |2     |
|2020-09-26|3       |4         |3      |1     |
|2020-09-27|7       |5         |4      |3     |
|2020-09-28|0       |2         |4      |6     |
|2020-09-29|2       |4         |3      |4     |
|2020-09-30|1       |2         |3      |4     |
|2022-02-01|1       |5         |5      |6     |
|2022-02-02|1       |6         |8      |4     |
|2022-02-03|1       |9         |9      |8     |
|2022-02-04|1       |4         |1      |2     |
|2022-02-05|1       |2         |2      |1     |
|2022-02-06|1       |4         |5      |4     |
|2022-02-07|1       |1         |1      |3     |
|2020-10-01|1       |2         |       |4     |
|2020-10-02|1       |2         |       |4     |
|2020-10-03|1       |2         |       |4     |
|2020-10-04|1       |2         |       |4     |
|...       |...     |...       |...    |...   |

I really need the code to be robust to the possibility that there may not be data for all dates for all countries. Can you guys help me?