Merging DataFrames with pandas（datacamp）

• Preparing data

  1.reading multiple data file

  #for loop
  df = [pd.read_csv(file) for file in filenames]
  
  #grob
  from glob import glob
  filenames = glob('___')   #包含通配符
  df = [pd.read_csv(file) for file in filenames]
  
  

  2.reindexing DataFrame

  indices: many index labels within index data structure

  indexes: many pandas index data structures

  df = df.reindex(___)
  df.sort_index()
  

  3.Arithmetic with Series&DataFrame

  df.divide(df1, axis='___')
  df.add(df1, fillvalue=0)
  df.pct_change() * 100
  df = df.multiply()
  

• Concatenating data

  1.appending and concatenating Series

  #append: stacks rows of df2 below s1
  Series1.append(Series2)
  
  #cancat can srack row_wise or column_wise:axis0 | axis=1
  pd.concat([Series1, Series2,...,Seriesn])
  

  2.appending and concatenating DataFrame

  df1.append(df2)
  pd.concat([df1, df2,...,dfn])
  

  3.concatenation, key, multilIndexes

  #how='inner'or how='outer'
  pd.concat()
  

  4.outer and inner joins

  np.hstack([___])
  np.vstack([___])
  np.concatenate([___], axis=___)
  pd.concat([___], keys=___, axis=___,join=___)
  

• Merging data

  1.merging DataFrames

  #how 'inner', 'outer','left', 'right'
  #many joins
  pd.merge(___, on=___, how=___)
  
  #merge_ordered:函数允许组合时间序列和其他有序数据。 特别是它有一个可选的fill_method关键字来填充/插入缺失的数据。
  pd.merge_ordered(___, on=___, suffixes=___, fill_method=___)
  
  #merge_asof:除了我们匹配最近的键而不是相等的键之外，其他的都类似于有序的left-join 。 对于左侧DataFrame中的每一行，我们选择右侧DataFrame中on键对应的值小于left的键对应的值的最后一行。
  

  2.joining

  #how 'inner', 'outer','left', 'right'
  pd.join(___, how=___)