def read_mapr03r_table(schema, table):
    '''
    Given the schema and table names in mapr03r, returns a Pandas
    DataFrame. This function piggybacks on Mac's Swagger tool
    (location: http://hdp0019.infores.com:9999/__swagger__/
    source: http://csmrnd01.infores.com:3000/CSM/r-apis/src/master/hive-meta/hive-meta-api.R)
    The significant advantage over the function `run_pyhive_query`
    (which reads a PyHive query and returns a pandas DataFrame)
    is the reduction in time. Code works for both text and ORC formats.
    Limitation: Currently works only for mapr03r tables.

    Parameters:
    ----------
    schema: schema name in mapr03r
    table: table name in the aforementioned schema

    Returns:
    ----------
    table as a Pandas DataFrame object
    '''
    from os.path import join
    from os import listdir
    from pyarrow import orc
    import pandas as pd
    import requests

    api_endpoint = 'http://hdp0019.infores.com:9999/describe'
    request_url = "%s?table=%s&schema=%s"%(api_endpoint, table, schema)
    resp = requests.get(request_url).json()

    col_names = [item['name'] for item in resp[0]]
    table_dir = [item['value'] for item in resp[1] if item['property'] == 'Location'][0]
    sep = [item['value'] for item in resp[1] if item['property'] == 'field.delim']
    if len(sep):
        sep = sep[0]
    else:
        sep = '|' # because some files don't have the field.delim attribute
    input_format = [item['value'] for item in resp[1] if item['property'] == 'InputFormat'][0]
    chunks = []

    if input_format == 'org.apache.hadoop.mapred.TextInputFormat':
        for file in listdir(table_dir):
            chunk = pd.read_csv(join(table_dir, file), sep=sep, header=None, names=col_names)
            chunks.append(chunk)
    elif input_format == 'org.apache.hadoop.hive.ql.io.orc.OrcInputFormat':
        for file in listdir(table_dir):
            filepath = join(table_dir, file)
            with open(filepath, 'rb') as f:
                data = orc.ORCFile(f) # pyarrow.Table object
                chunk = data.read().to_pandas()
            chunks.append(chunk)
    else:
        print('Incorrect input format, aborting read.')
        return None
    return pd.concat(chunks)

def run_pyhive_query(server, query, drop_schema_name=True):
    '''
    Runs a query in one of the hadoop clusters, mapr03r or mapre04p.
    Since this authenticates the connection using netrc, one
    needs to write a file ~/.netrc resembling the following
    ```machine hdp0019.infores.com login msabp password *****```
    Appropriate permissions (chmod 666) should be set to this file.

    Parameters:
    ----------
    server: `mapr03r` or `mapre04p`
    query: SQL query in three ticks, ''' '''
    split_dot: bool (default True), if True drops the schema name from
            DataFrame generated by query

    Returns:
    ----------
    pd.DataFrame object with database name dropped from columns.
    '''

    import pandas as pd
    from pyhive import hive
    from netrc import netrc

    if server == 'mapr03r':
        host = "hdp0019.infores.com"
    elif server == 'mapre04p':
        host = "lnx1170.ch3.prod.i.com"
    else:
        print("Server name not entered correctly!")
        return None

    info = netrc()
    user = info.authenticators(host)[0]
    passwd = info.authenticators(host)[2]
    auth = "LDAP"
    port = "10000"

    # establish connection
    conn = hive.Connection(host=host, port=port, username=user, password=passwd, auth=auth)
    try:
        df = pd.read_sql(query, conn)
        if drop_schema_name:
            if all([item.count('.')==1 for item in df.columns]):
                df.columns = [item.split(".")[1] for item in df.columns]
            else:
                print("Cannot drop schema name from query.")
        return df
    except TypeError:
        print("TypeError: 'NoneType' object is not iterable, since query does not return table!")
        return None