8 examples of 'merge in pandas' in Python

29
def _concat(df, type):
30
    if df is None:
31
        df = pd.DataFrame(_object_blocks[type])
32
    else:
33
        _df = pd.DataFrame(_object_blocks[type])
34
        df = pd.concat([df, _df], sort=True)
35
    return df

332
def merge_datasets(self, other):
333
    """
334
    This operation combines two dataframes into one new DataFrame.
335
    If the operation is combining two SpatialDataFrames, the
336
    geometry_type must match.
337
    """
338
    if isinstance(other, SpatialDataFrame) and \
339
        other.geometry_type == self.geometry_type:
340
        return pd.concat(objs=[self, other], axis=0)
341
    elif isinstance(other, DataFrame):
342
        return pd.concat(objs=[self, other], axis=0)
343
    elif isinstance(other, Series):
344
        self['merged_datasets'] = other
345
    elif isinstance(other, SpatialDataFrame) and \
346
         other.geometry_type != self.geometry_type:
347
        raise ValueError("Spatial DataFrames must have the same geometry type.")
348
    else:
349
        raise ValueError("Merge datasets cannot merge types %s" % type(other))

6
def default_merger(x, y):
7
	import pandas as pd
8
	return pd.concat([x, y])

813
def merge(old_cols, new_cols):
814
    return old_cols + new_cols

142
def append_data(df1, df2):
143
    '''
144
    Append df2 to df1
145
    '''
146
    df = pd.concat((df1, df2))
147
    return df.groupby(df.index).first()

51
def random_merge(A, B, N=20, on='AnswerId', key='key', n='n'):
52
    """Pair all rows of A with 1 matching row on "on" and N-1 random rows from B
53
    """
54
    assert key not in A and key not in B
55
    X = A.copy()
56
    X[key] = A[on]
57
    Y = B.copy()
58
    Y[key] = B[on]
59
    match = X.merge(Y, on=key).drop(key, axis=1)
60
    match[n] = 0
61
    df_list = [match]
62
    for i in A.index:
63
        X = A.loc[[i]]
64
        Y = B[B[on] != X[on].iloc[0]].sample(N-1)
65
        X[key] = 1
66
        Y[key] = 1
67
        Z = X.merge(Y, how='outer', on=key).drop(key, axis=1)
68
        Z[n] = range(1, N)
69
        df_list.append(Z)
70
    df = pd.concat(df_list, ignore_index=True)
71
    return df

474
def full_merge_of_existing_series(old_series, new_series):
475
    """
476
    Merges old data with new data.
477
    Any Nan in the existing data will be replaced (be careful!)
478
479
    :param old_data: pd.Series
480
    :param new_data: pd.Series
481
482
    :returns: pd.Series
483
    """
484
    if len(old_series)==0:
485
        return  new_series
486
    if len(new_series)==0:
487
        return old_series
488
489
    joint_data = pd.concat([old_series, new_series], axis=1)
490
    joint_data.columns = ['original', 'new']
491
492
    # fill to the left
493
    joint_data_filled_across = joint_data.bfill(1)
494
    merged_data = joint_data_filled_across['original']
495
496
    return merged_data

88
def merge_lvj_dfs(df1, df2, how='outer'):
89
    """
90
    Merge two data frames on lvj indices.
91
92
    By default, uses the union of the keys (an "outer" join).
93
    """
94
    merged = pd.merge(df1, df2, how=how, left_index=True, right_index=True)
95
    merged.fillna(0, inplace=True)
96
    return merged