Functions
| Task | Code |
|---|---|
| Read file | pd.read_csv(path) |
| Set maximum rows displayed in the terminal | pd.options.display.max_rows: number |
| Recognize headers | print(df.columns) |
| Display particular headers | print(df[['Categories']]) |
| Sort values alphabetically | df.sort_values(['Search Volume'], ascending: True).head(100)Use ascending=False for descending values |
| Values in particular condition | print(df.loc[(df['Intents'] == 't') | | Drop column |df.drop(df.columns[[1, 2, 3, 4]], axis: 1)| | Insert a new column |df[‘new category’] = df.iloc[:, range].sum(axis: 1)| | Rearrange headers |list(df.columns.values)| | Save as CSV |df.to_csv(‘path’, index: False)<br>Usesepparameter for a different separator | | Save as XLS |df.to_excel(‘path’, index: False)| | Mean |df[‘category’].mean()| | Trimmed mean |trim_mean(df[‘category’], drop_percentage)<br>Drop percentage of values from top and bottom | | Median |df[‘category’].median()| | Standard deviation |df[‘category’].std()| | Check duplicated values |df.loc[df.duplicated()]<br>Usesubset=[name_of_the_column]to check a specific column | | Duplicated values |df.duplicated()<br>The output is a boolean Series. Usedf.duplicated(subset=[name_of_the_column])to check a specific column | | Count of NaN values |df.isna().sum()| | Rename columns |df.rename(columns={‘previous_column’: ‘new_column’})| | Count of each unique value in a column |df[‘MAIN_GENRE’].value_counts()| | Concatenate two data frames (append) |pd.concat([df, df_to_append])| | Convert to numeric (float or int) |pd.to_numeric(df[‘numeric category’])| | Convert to datetime format |pd.to_datetime(df[‘publishTime’])| | Casting data types |df[‘column’].astype(‘data_type’)| | Filter out rows with no values |df.loc[\~df[‘likeCount’].isna()]| | Check for NaN values |df.isna()<br>It returns a boolean DataFrame. Use\~before the expression to get the positive values | | Subset rows based on condition |df.loc[df[‘RELEASE_YEAR’] > 1999]<br>You can also use thequerymethod | | Subset columns |df[[‘column1’, ‘column2’]]| | Shape of the DataFrame |df.shape| | Basic information |df.describe()| | Set the index |df.set_index(‘column_name’)| | Locate based on index |df.loc[index_value]| | Data types |df.dtypes`Remember not to use parentheses |
| Get the only the values of the column in a list of lisit | df[’’[‘Title’]].values.tolist() |