How to preserve the datatype while iterating dataframe in pandas?

Question

If I print out a dataframe directly, I get the correct output with correct datatypes. However, when I try to iterate the same dataframe, the datatypes are changing.

Here is my program:

import pandas as pd

F = 9.37556366342
p = 0.000101673198518
df_between = 2
df_within = 471
df_total = 473

summary_stats_vals = [(F,p,df_between,df_within,df_total)]
labels = ['F-statistics', 'p-value', 'df-between', 'df-within', 'df-total']
df = pd.DataFrame.from_records(summary_stats_vals,columns=labels)

print(df)
print()

#Iterating the dataframe
for index, row in df.iterrows():
    df_row = list()
    df_row.append(index)
    for col in df.columns:
        df_row.append(row[col])
    print(row)

As you can see from the screenshot below, the data types of df_between, df_within and df_total have not been preserved while iterating. They are changing from int to float data type. What is the way I can preserve the data types while iterating a dataframe?

Answer 1

From the docs:

Because iterrows returns a Series for each row, it does not preserve dtypes across the rows (dtypes are preserved across columns for DataFrames).

You could use DataFrame.itertuples() and get namedtuples for each row.

>>> for r in df.itertuples(index=False):
...     print(r)

Pandas(_0=9.3755636634199995, _1=0.000101673198518, _2=2, _3=471, _4=473)
>>> for r in df.itertuples(index=False):
...     print(r._3)

471

Changing your column names to valid Python identifiers might make more sense:

...
labels = ['F_statistics', 'p_value', 'df_between', 'df_within', 'df_total']
...

>>> for r in df.itertuples(index=False, name='Stuff'):
...     print(r)

Stuff(F_statistics=9.3755636634199995, p_value=0.000101673198518, df_between=2, df_within=471, df_total=473)
>>>
>>> for r in df.itertuples(index=False, name='Stuff'):
...     print(r.df_total)

473
>>> 

---

I haven't found, in the docs, an explicit statement that Series datatype is homogeneous, but it is inferred,it acts like a Numpy ndarray, and the constructor has a dtype parameter which applies to all the values in the Series:

One-dimensional ndarray with axis labels (including time series).

Looks like even if only one value in the Series is a float, the series dtype will be float:

>>> s = pd.Series([1,2,3,4.1], index=['a','b','c','d'])
>>> s
a    1.0
b    2.0
c    3.0
d    4.1
dtype: float64
>>> 

Answer 2

Thank you so much wwii. Yeah that worked out very well. The code below is what I needed. Thanks again for your help.

for r in df.itertuples(index=False, name='summary_stats'):
    for item in r:
        print(item)

I get this output:

>>>9.37556366342
   0.000101673198518
   2
   471
   473