Python useful functions

Python_helper.txt

-------------------
Pandas
-------------------

# Create dataframe from dictionary
df = pd.DataFrame.from_dict(data["trainData"])

# Convert entire dataframe values to numeric type.
df = df.apply(pd.to_numeric, errors='ignore')

# Remove missing value.
df = df.dropna()

# Create a new column named "new_col" with value "2" if cluster_id column has value 1, else 0.
df.insert(loc=0, column=new_col, value=np.where(cluster_id == 1, 2, 0))

# Create a new column named "new_col" with value of df["log2_livable"] if cluster_id is 1, else 0.
df.insert(loc=0, column=new_col, value=np.where(cluster_id == 1, df["log2_livable"], 0))

# Change date format of entire column.
effective_date = pd.to_datetime(df["effective_date"], format='%Y-%m-%d')

# Re-order dataframe by columns (A-Z)
df = df[sorted(df.columns)]

# Clone dataframe
df.copy(deep=True)

# Fetch the row based on given index.
df.iloc[[index]]

# Fetch multiple columns.
df.loc[:, ['col_1', 'col_2']]

# Filter rows in dataframe where cluster_2 is 0
X[X['cluster_2'].isin([0])].reset_index()

# Sum of rows
df = df.sum(axis = 1)

# Create dataframe from dictionary (Key as columns)
a={'b':100,'c':300}
pd.DataFrame(coefs, index=[0,])