top of page
Search

Pandas Part 6

How to fill null values

  • Null (nan) values can be filled by using fillna() method (DataFrame.fillna())

  • It can be applied to series as well as for data frame objects

Example 1

** all the nan have been filled with 0

List of Important Parameters fillna() takes

  • value ---- "which value to be filled "

  • method ---- "different methods to fill nan "

  • axis ---- " Axis along which to fill missing values" [0 or ‘index’, 1 or ‘columns]

  • inplace ---- "True or False " default will be False. If True, do operation inplace and return None

  • limit ---- "To limit the fill for methods used to fill nan "



value

  • which values to be filled In nan position

  • If a single value is passed then all the nan will be replaced with that single value passed

Example 2

  • we can also pass dictionaries where keys will indicate the column name and values indicating which value to replace with nan in that particular column

Example 3

* values passed should be logical to that column otherwise total column datatype will get effected

Example:- in eps we can fill nan with mean of that column and in tickers we can fill with the mode of that column


inplace

  • When ever we are using fillna method on the data structures it is not affecting the original data structures .It is returning a new data structure object

  • if we want to modified the data in place,we have to use inplace = True which means it will return nothing and the dataframe is now updated.

  • If inplace = False which means it will return new dataframe and the dataframe is now updated

Example 4

Example 5


method

  • We also use some methods to fill na values

  • We can use "ffill" method knows as forward fill .fill is used to forward fill the missing values in the dataset

Example 6

** in 2 index row in "pe" column the nan values is filled with 1 index value in "pe" column (forward fill)


  • We can use "bfill" method knows as backward fill .bfill is used to backward fill the missing values in the dataset

Example 7

** in 2 index row in "pe" column the nan values is filled with 3 index value in "pe" column (backward fill)


axis

  • While using method parameters to fill nan values we can also use axis to specify which axis to consider whether column or rows to fill nan values

  • axis=0 means column axis=1 means rows

Example 8

Example 9

** when axis = 1 it is doing ffill row wise thats why in "pe" column in 2 index the nan value has filled the values from 2 index form "eps" column


limit

  • limit parameter is used to limit the maximum number of consecutive NaN values to be filled by the method.

Example 10

** all the consecutive nan values are filled

Example 11

** only the first consecutive nan value are filled

Example 12

** only the two consecutive nan values are filled


How to drop null values

  • Null (nan) values can be droped by using dropna() method (DataFrame.dropna())

  • It can be applied to series as well as for data frame objects

  • It will drop the rows or columns containing nan values

  • Default it will drop rows

Example 13

** dropped 2,3,4,6 indexed rows (because all this rows contain nan values


List of Important Parameters dropna() takes

  • how ---- "how to drop the rows "

  • axis ---- "columns or rows to drop"

  • inplace ---- "True or False " default will be False. If True, do operation inplace and return None

  • thresh ---- "Require that many non-NA values "


how

  • how to drop the rows

  • how parameters take two argument "all" or" any"

Example 14

  • "all" -- If all values are NA, drop that row or columns

  • "any" -- If any values are NA, drop that row or columns

Example 15


axis

  • Here also we can apply axis to specify which axis (column or rows) to drop if it contains nan values

Example 16

** axis = 0 drop rows containing nan values

Example 17

** axis = 1 drop columns containing nan values


inplace

  • When ever we are using fillna method on the data structures it is not affecting the original data structures .It is returning a new data structure object

  • if we want to modified the data in place,we have to use inplace = True which means it will return nothing and the dataframe is now updated.

  • If inplace = False which means it will return new dataframe and the dataframe is now updated

Example 18


thresh

  • How many non nan values are required to remove rows or columns

  • Here we are dealing with non nan values not nan values

Example 19

** drop those rows without 1 non nan values

Example 20

** drop those rows without 3 non nan values

Example 21

** drop those rows without 4 non nan values


33 views0 comments

Related Posts

Comentários


bottom of page