An evolving set of pandas snippets I find useful

pandas.py

# Unique values in a dataframe column
df['column_name'].unique()

# Grab dataframe rows where column = value
df = df.loc[df.column == 'some_value']

# Grab dataframe rows where column value is present in a list
value_list = ['value1', 'value2', 'value3']
df = df.loc[:,df.columns.isin(valuelist)]
# or grab rows where a value is not present in a list
df = df.loc[:,~df.columns.isin(valuelist)]

# Delete column from dataframe
del df['column_name']

# Lower-case all dataframe column names
df.columns = map(str.lower, df.columns)

# Lower-case everything in a dataframe column
df.column_name = df.column_name.str.lower()

# Get the length of data in a dataframe column
df.column_name.str.len()

# Get a quick count of rows in the dataframe
len(df.index)

# Concatenate two dataframe columns into a new column
df['new_column'] = df['column1'].astype(str) + df['column2'].astype(str)

# Create a dataframe from a Python dictionary
df = pd.DataFrame(list(a_dictionary.items()), columsn = ['column1', 'column2']