pandas.merge合并

Help on function merge in module pandas.core.reshape.merge:

merge(left, right, how='inner', on=None, left_on=None, right_on=None, left_index=False, right_index=False, sort=False, suffixes=('_x', '_y'), copy=True, indicator=False)

    Merge DataFrame objects by performing a database-style join operation by

    columns or indexes.

    

    If joining columns on columns, the DataFrame indexes *will be

    ignored*. Otherwise if joining indexes on indexes or indexes on a column or

    columns, the index will be passed on.

    

    Parameters

    ----------

    left : DataFrame

    right : DataFrame

    how : {'left', 'right', 'outer', 'inner'}, default 'inner'

        * left: use only keys from left frame, similar to a SQL left outer join;

          preserve key order只保留左表的所有数据        

     * right: use only keys from right frame, similar to a SQL right outer join;

          preserve key order只保留右表的所有数据

      * outer: use union of keys from both frames, similar to a SQL full outer

          join; sort keys lexicographically保留两个表的所有信息

     * inner: use intersection of keys from both frames, similar to a SQL inner

          join; preserve the order of the left keys只保留两个表中公共部分的信息

  on : label or list

        Field names to join on. Must be found in both DataFrames. If on is

        None and not merging on indexes, then it merges on the intersection of

        the columns by default.

    left_on : label or list, or array-like

        Field names to join on in left DataFrame. Can be a vector or list of

        vectors of the length of the DataFrame to use a particular vector as

        the join key instead of columns

    right_on : label or list, or array-like

        Field names to join on in right DataFrame or vector/list of vectors per

        left_on docs

    left_index : boolean, default False

        Use the index from the left DataFrame as the join key(s). If it is a

        MultiIndex, the number of keys in the other DataFrame (either the index

        or a number of columns) must match the number of levels

    right_index : boolean, default False

        Use the index from the right DataFrame as the join key. Same caveats as

        left_index

    sort : boolean, default False

        Sort the join keys lexicographically in the result DataFrame. If False,

        the order of the join keys depends on the join type (how keyword)

    suffixes : 2-length sequence (tuple, list, ...)

        Suffix to apply to overlapping column names in the left and right

        side, respectively

    copy : boolean, default True

        If False, do not copy data unnecessarily

    indicator : boolean or string, default False

        If True, adds a column to output DataFrame called "_merge" with

        information on the source of each row.

        If string, column with information on source of each row will be added to

        output DataFrame, and column will be named value of string.

        Information column is Categorical-type and takes on a value of "left_only"

        for observations whose merge key only appears in 'left' DataFrame,

        "right_only" for observations whose merge key only appears in 'right'

        DataFrame, and "both" if the observation's merge key is found in both.

    

        .. versionadded:: 0.17.0

    

    Examples

    --------

    

    >>> A              >>> B

        lkey value         rkey value

    0   foo  1         0   foo  5

    1   bar  2         1   bar  6

    2   baz  3         2   qux  7

    3   foo  4         3   bar  8

    

    >>> A.merge(B, left_on='lkey', right_on='rkey', how='outer')

       lkey  value_x  rkey  value_y

    0  foo   1        foo   5

    1  foo   4        foo   5

    2  bar   2        bar   6

    3  bar   2        bar   8

    4  baz   3        NaN   NaN

    5  NaN   NaN      qux   7

    

    Returns

    -------

    merged : DataFrame

        The output type will the be same as 'left', if it is a subclass

        of DataFrame.

    

    See also

    --------

    merge_ordered

    merge_asof