dataframe_example.py

import pandas as pd
import numpy as np
from datetime import datetime

# generate some fake tick data with 1 million observations
n = 1000000
df = pd.DataFrame({
    "timestamp": [datetime.now() for t in range(n)],
    "value": np.random.uniform(-1, 1, n)
    })

# similar dataframe operations to R
df.head()
df.describe()
df.count()

# timing a basic operation on the data frame
s = time.time(); df.describe(); e = time.time()-s;
print e
# 0.7173671722412109

# creating an index
df['timestamp'] = pd.to_datetime(df['timestamp'])
df['value'] = df['value'].cumsum()
df.index = df['timestamp']

# index improves performance
s = time.time(); df.describe(); e = time.time()-s;
print e
# 0.5612149238586426