Loading Data

• import pandas as pd
• tables are called data frames (df)
• read csv: pd.read_csv()
  • arguments
    • na_values
• view column
  • df['col_name']
• view unique values: np.unique()
• drop rows with missing values: df.dropna()
• see column names: df.columns
• accessing rows and columns is similar to an array
• select specific columns: df[['col1','col2']]
• see current index: df.index
  • re-name the rows based on contents in a column: df.set_index('col')
    • select rows based on new contents: df.loc['value']
• select rows by numeric position: df.iloc[[3,4]]
• select columns by numeric position: df.iloc[:,[0,2,3]]
• select specific rows and columns: df.iloc[[3,4],[0,2,3]]
• index entries don't need to be unique
• lambda use inside of function calls to return single expressions
  • i.e. Auto_re.loc[lambda df: df['year'] > 80, ['weight', 'origin']]
• & is 'and'
• | is 'or'
• For integer based queries, use the iloc[] method. For string and Boolean selections, use the loc[] method. For functional queries that filter rows, use the loc[] method with a function (typically a lambda) in the rows argument.
• access columns as an attribute: df.col
• bin data by using pd.cut()
• drop data by using pd.drop()