pyspark.sql.functions.try_divide#

pyspark.sql.functions.try_divide(left, right)[source]#

Returns dividend/divisor. It always performs floating point division. Its result is always null if divisor is 0.

Added in version 3.5.0.

Parameters:
leftColumn or column name

dividend. A column that evaluates to a numeric or interval.

rightColumn or column name

divisor. A column that evaluates to a numeric.

Examples

Example 1: Integer divided by Integer.

>>> import pyspark.sql.functions as sf
>>> spark.createDataFrame(
...     [(6000, 15), (1990, 2), (1234, 0)], ["a", "b"]
... ).select("*", sf.try_divide("a", "b")).show()
+----+---+----------------+
|   a|  b|try_divide(a, b)|
+----+---+----------------+
|6000| 15|           400.0|
|1990|  2|           995.0|
|1234|  0|            NULL|
+----+---+----------------+

Example 2: Interval divided by Integer.

>>> import pyspark.sql.functions as sf
>>> df = spark.range(4).select(sf.make_interval(sf.lit(1)).alias("itvl"), "id")
>>> df.select("*", sf.try_divide("itvl", "id")).show()
+-------+---+--------------------+
|   itvl| id|try_divide(itvl, id)|
+-------+---+--------------------+
|1 years|  0|                NULL|
|1 years|  1|             1 years|
|1 years|  2|            6 months|
|1 years|  3|            4 months|
+-------+---+--------------------+

Example 3: Exception during division, resulting in NULL when ANSI mode is on

>>> import pyspark.sql.functions as sf
>>> origin = spark.conf.get("spark.sql.ansi.enabled")
>>> spark.conf.set("spark.sql.ansi.enabled", "true")
>>> try:
...     spark.range(1).select(sf.try_divide("id", sf.lit(0))).show()
... finally:
...     spark.conf.set("spark.sql.ansi.enabled", origin)
+-----------------+
|try_divide(id, 0)|
+-----------------+
|             NULL|
+-----------------+