concat_row_vs_df.py

import pandas as pd
import numpy as np

def add_individual_rows():
    heights = range(150, 211)
    satisfaction_baseline = 150
    d = {}
    l = []
    df = pd.DataFrame(columns=['height', 'satisfaction'])
    for h in heights:
        loc = satisfaction_baseline*((h/100.)**0.5)
        size = np.random.randint(20,30)
        dist = np.random.normal(loc=loc, scale=1.0, size=size)
        d[h] = dist
        l.append((h, dist))
        
        for i in dist:
            df = df.append({'height': h, 'satisfaction': i}, ignore_index=True)
    return df
  
def add_multiple_rows():
    heights = range(150, 211)
    satisfaction_baseline = 150
    d = {}
    l = []
    df = pd.DataFrame(columns=['height', 'satisfaction'])
    for h in heights:
        loc = satisfaction_baseline*((h/100.)**0.5)
        size = np.random.randint(20,30)
        dist = np.random.normal(loc=loc, scale=1.0, size=size)
        d[h] = dist
        l.append((h, dist))
        rows = [(h, i) for i in dist]
        temp = pd.DataFrame(rows, columns=['height', 'satisfaction'])
        df = pd.concat([df, temp])
    return df
  
  
%%time
x = add_individual_rows()
#Wall time: 4.23 s

%%time
y = add_multiple_rows()
#Wall time: 179 ms