Hi, I came across this strange behavior of Apache Spark 1.6.1: when I was reading mysql table into spark dataframe ,a column of data type float got mapped into double.
dataframe schema: root |-- id: long (nullable = true) |-- ctime: double (nullable = true) |-- atime: double (nullable = true) mysql schema: mysql> desc test.user_action_2; +-------+------------------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +-------+------------------+------+-----+---------+-------+ | id | int(10) unsigned | YES | | NULL | | | ctime | float | YES | | NULL | | | atime | double | YES | | NULL | | +-------+------------------+------+-----+---------+-------+ I wonder if anyone have seen this behavior before.