create_inspections_subset.py 文件源码

def create_subset(src, dest, n=250):
    "Given a csv file `src`, create a subset `dest` with `n` unique entities"
    df = pd.read_csv(src)
    lics = pd.unique(df["License #"])
    sublics = lics[random.sample(range(0,len(lics)), n)]
    subset = df[df["License #"].isin(sublics)]
    # Make the column names a little more readable
    subset.columns = map(clean_column_name, subset.columns)
    subset.to_csv(dest, index=False)