Pandas: Transpose the first 3 rows but duplicate the rest of the data so I get a unique row and a larger table - python-3.x

Question

I'm trying to convert an existing excel sheet that has 3 layers of columns. The first column is the year, but it's a merged cell. The 2nd column is of the months, also merged, and the 3rd layer is alternating Rent | Other.
Original data is shaped like this:
My data looks like this:
2022
Unnamed: 1
Unnamed: 2
Unnamed: 3 ...
2023
Unnamed: 135
Unnamed: 136 ...
January
NaN
February
NaN
January
NaN
February
Rent
Other
Rent
Other
Rent
Other
Rent
100
0
120
30
110
25
100
I added the "...." to the table, this continues for ~130 or so columns per year.
I tried to forward fill the year and months:
2022
2022
2022
2022
January
January
February
February
Rent
Other
Rent
Other
100
0
120
30
I want it to look like this:
Year
Month
Rent
Other
2022
January
100
0
2022
February
120
30

Answer 1

Flying blind here since I don't have access to your Excel file.
df = (
# Your file look to be row-oriented instead of pandas' usual column-oriented
# format. We will import it without column names and assign them later.
pd.read_excel("file.xlsx", header=None)
# Fill in the blanks since some of the cells are merged
.ffill(axis=1)
# Set the row's index, then transpose the dataframe to the usual
# column-oriented format
.set_axis(["Year", "Month", "Metric", "Value"])
.T
)
# Month name is usually a pain in the neck to work with. By default, they sort
# in alphabetical order so April, August, February, ... It's best to convert
# them into number, but if you want to keep the name, use CategoricalDType to
# keep them in semantic order
MonthDType = pd.CategoricalDtype(
pd.date_range("2022-01-01", "2022-12-01", freq="M").strftime("%B"), ordered=True
)
df["Month"] = df["Month"].astype(MonthDType)
# The final pivot
df = df.pivot_table(
index=["Year", "Month"],
columns="Metric",
values="Value",
aggfunc="sum",
observed=True,
).reset_index()