import pandas as pd

df = pd.read_csv('data.txt',names=['id', 'os', 'dim', 'uv', 'pv'],sep=',')

df

import numpy as np

row_df = pd.DataFrame(np.array([['2', 'ios', '苹果-iPad 4', 3287509, 32891811]]), columns=['id', 'os', 'dim', 'uv', 'pv'])
df = df.append(row_df, ignore_index=True)

df

df['time'] = '2016-07-19'
df

def where(df, column_name, id_value):
    df = df[df[column_name] == id_value]
    return df


def to_dict(df):
    """
    {"pv" or "uv" -> {"os": os_value}}
    :return: dict
    """
    df = where(df, 'id', 0)
    df_dict = df.set_index('os')[['uv', 'pv']].to_dict()
    return df_dict

def group_top(df, group_col, sort_col, top_n):
    """
    get top(`sort_col`) after group by `group_col`
    :param df: dataframe
    :param group_col: string, column name
    :param sort_col: string, column name
    :param top_n: int
    :return: dataframe
    """
    return df.assign(rn=df.sort_values([sort_col], ascending=False)
                     .groupby(group_col)
                     .cumcount() + 1) \
        .query('rn < ' + str(top_n + 1)) \
        .sort_values([group_col, 'rn'])
        
def top(df, group_col, sort_col, top_n):
    """overall top and group top"""
    all_top_df = df.nlargest(top_n, columns=sort_col)
    grouped_top_df = group_top(df, group_col, sort_col, top_n)
    grouped_top_df = grouped_top_df.ix[:, 0:-1]
    result_df = pd.concat([all_top_df, grouped_top_df]).drop_duplicates()
    return result_df

df['rank'] = df['uv'].rank(method='first', ascending=False).apply(lambda x: int(x))
df

def left_join(left, right, on, right_col, default_value):
    df = pd.merge(left, right, how='left', on=on)
    df[right_col] = df[right_col].map(lambda x: default_value if pd.isnull(x) else x)
    return df

def percentage(part, whole):
    return round(100*float(part)/float(whole), 2)


os_dict = to_dict(df)
all_uv = sum(os_dict['uv'].values())
df = where(df, 'id', 1)
df['per'] = df.apply(lambda r: percentage(r['uv'], all_uv), axis=1)
df

D:\Anaconda2\envs\py3\lib\site-packages\ipykernel_launcher.py:8: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
  

duplicate = df.duplicated(subset=['os','rank'], keep=False)
duplicate

2     True
3     True
4     True
5     True
6     True
7    False
dtype: bool