To easily share a sample dataframe using df.to_dict() with Python Pandas, we can split a dataframe with 'split' with calling to_dict.
For instance, we write
import plotly.express as px
df = px.data.iris().head(10)
df.to_dict('split')
to call to_dict on our dataframe with 'split' to return a data frame that has some of the rows that’s in the data frame.
To drop rows of a Python Pandas DataFrame whose value in a certain column is NaN, we call the notna method.
For instance, we write
df = df[df['EPS'].notna()]
to filter out the rows with 'EPS' column value that’s NaN by calling thw notna method on the column.
Then we get the filtered rows with df[df['EPS'].notna()]
To do fuzzy match merge with Python Pandas, we can use the fuzzymatcher library.
To install it, we run
pip install fuzzymatcher
Then we use it by writing
from fuzzymatcher import link_table, fuzzy_left_join
df1 = pd.DataFrame({'Col1':['Microsoft', 'Google', 'Amazon', 'IBM']})
df2 = pd.DataFrame({'Col2':['Mcrsoft', 'gogle', 'Amason', 'BIM']})
left_on = ["Col1", "Col2"]
right_on = ["Col2", "Col2"]
fuzzymatcher.link_table(df1, df2, left_on, right_on)
to create 2 dataframes df1 and df2.
Then we call the fuzzymatcher.link_table method to merge df1 and df2 on the columns listed in left_on and right_on.
To select DataFrame rows between two dates with Python Pandas, we can use a boolean mask.
For instance, we write
df['date'] = pd.to_datetime(df['date'])
mask = (df['date'] > start_date) & (df['date'] <= end_date)
to convert the 'date' column entries to datetime64 with pd.to_datetimr.
Then we create the mask with (df['date'] > start_date) & (df['date'] <= end_date).
And then we get the filtered rows between start_date and end_date with
df.loc[mask]
To do conditional replace with Python Pandas, we can return the items we want to replace and then replace them.
For instance, we write
mask = df.my_channel > 20000
column_name = 'my_channel'
df.loc[mask, column_name] = 0
to replace them items returned by mask, which is the rows with my_channel column values that are bigger than 20000.
Then we use
df.loc[mask, column_name] = 0
to replace the returned rows and columns with .