#!/usr/bin/env python
# coding: utf-8

# # How do I use the MultiIndex in pandas? ([video](https://www.youtube.com/watch?v=tcRGa2soc-c&list=PL5-da3qGB5ICCsgW1MxlZ0Hq8LL5U3u9y&index=31))
# 
# - [My pandas video series](https://www.dataschool.io/easier-data-analysis-with-pandas/)
# - [GitHub repository](https://github.com/justmarkham/pandas-videos)

# In[1]:


import pandas as pd


# In[2]:


stocks = pd.read_csv('data/stocks.csv')
stocks


# In[3]:


stocks.index


# In[4]:


stocks.groupby('Symbol').Close.mean()


# ## Series with MultiIndex

# In[5]:


ser = stocks.groupby(['Symbol', 'Date']).Close.mean()
ser


# In[6]:


ser.index


# In[7]:


ser.unstack()


# In[8]:


df = stocks.pivot_table(values='Close', index='Symbol', columns='Date')
df


# ## Selection from Series with MultiIndex

# In[9]:


ser


# In[10]:


ser.loc['AAPL']


# In[11]:


ser.loc['AAPL', '2016-10-03']


# In[12]:


ser.loc[:, '2016-10-03']


# In[13]:


df


# In[14]:


df.loc['AAPL']


# In[15]:


df.loc['AAPL', '2016-10-03']


# In[16]:


df.loc[:, '2016-10-03']


# ## DataFrame with MultiIndex

# In[17]:


stocks.set_index(['Symbol', 'Date'], inplace=True)
stocks


# In[18]:


stocks.index


# In[19]:


stocks.sort_index(inplace=True)
stocks


# ## Selection from DataFrame with MultiIndex

# In[20]:


stocks.loc['AAPL']


# In[21]:


stocks.loc[('AAPL', '2016-10-03'), :]


# In[22]:


stocks.loc[('AAPL', '2016-10-03'), 'Close']


# In[23]:


stocks.loc[['AAPL', 'MSFT'], :]


# In[24]:


stocks.loc[(['AAPL', 'MSFT'], '2016-10-03'), :]


# In[25]:


stocks.loc[(['AAPL', 'MSFT'], '2016-10-03'), 'Close']


# In[26]:


stocks.loc[('AAPL', ['2016-10-03', '2016-10-04']), 'Close']


# In[27]:


stocks.loc[(slice(None), ['2016-10-03', '2016-10-04']), :]


# ## Merging DataFrames with MultiIndexes

# In[28]:


close = pd.read_csv('data/stocks.csv', usecols=[0, 1, 3], index_col=['Symbol', 'Date']).sort_index()
close


# In[29]:


volume = pd.read_csv('data/stocks.csv', usecols=[0, 2, 3], index_col=['Symbol', 'Date']).sort_index()
volume


# In[30]:


both = pd.merge(close, volume, left_index=True, right_index=True)
both


# In[31]:


both.reset_index()