import numpy as np
import pandas as pd

arr = np.arange(4).astype('float64')
arr[2] = np.nan

def sample_stdev(arr):
    # convert to dataframe
    df = pd.DataFrame(arr)
    # drop na samples
    df = df.dropna()
    # return std dev, just it's values, not the full df
    return df.std().value

def sample_stdev2(arr):
    arr = arr[-np.isnan(arr)]
    results = np.std(arr)
    return results

x = np.array([np.nan, 1, 2])
sample_stdev(x)
sample_stdev2(x)

from pandas import Series, DataFrame

goog = returns.ix[:, "GOOG"]
goog = returns.GOOG

val = [1, 2, 3]
labels = ['A', 'B', 'C']
s = Series(val, labels)

s.ix['A':'B']
s.values

# series is one dimension of a DataFrame, like a vector

val = [np.nan, 1, 2]
s = Series(val, labels)
s.std()
s.mean()
s.max()

# let's make a DataFrame
val = [[1, 2, 3],
        [4, 5, 6],
        [7, 8, 9]]
rows = ['B', 'C', 'D']
cols = ['X', 'Y', 'Z']
df = DataFrame(val, rows, cols)

df.ix['C', 'Y']
df.ix[['C', 'B'], 'Y']
type(df.ix[['C', 'B'], 'Y'])
df.ix[['C', 'B'], ['X', 'Y']]
df.index
df.columns
df.values

val = [1, 2, 3]
s = Series(val, labels)
s2 = df.ix[:, 'Y']
s + s2
s.add(s2, fill_value=0)

df.sum()
df.sum(axis=1)
df.mean()
df.std()

def cond_mean(returns):
    # avg daily returns of GOOG when AAPL is UP
    # and also when AAPL is DOWN
    ret1 = returns.ix[:, 0]
    ret2 = returns.ix[:, 1]
    mask = ret2 > 0
    up1 = ret1[mask]
    down1 = ret1[-mask]
    s = Series([up1.mean(), down1.mean()], ['UP', 'DOWN'])
    return s
    
rs = cond_mean(returns)
assert type(rs) == Series

val = [number, other_number]
Series(Val, ['UP', 'DOWN']