dropna

Dropna is a method in the pandas library used to remove missing values from DataFrame or Series objects. It can return a new object or modify the original one in place, depending on the inplace parameter. Dropna is a common data cleaning step to prepare data for analysis or modeling.

By default, dropna operates along rows (axis=0) and removes any row containing at least one missing value. Setting axis=1 switches to removing columns that contain any missing values. This makes it straightforward to exclude incomplete observations or features from a dataset.

The how parameter controls the criterion for removal: 'any' (the default) drops a row or column if any values are missing, while 'all' drops only if all values in the row or column are missing. The thresh parameter specifies the minimum number of non-missing values required to keep a row or column. The subset parameter allows evaluating missing values only within a specified list of columns, which can be useful when only certain features are considered for completeness.

In addition, ignore_index can be used to reindex the result with a continuous index, and inplace determines whether the operation mutates the original object or returns a modified copy. Practical use of dropna includes removing rows with missing critical data or excluding unreliable features, but it is important to consider the impact on dataset size and study design. Alternatives such as fillna for imputation or more advanced imputation strategies may be appropriate depending on the context. Related functions include isna and notna for detecting missing values and fillna for substituting them.