pyspark.sql.functions.try_subtract¶

pyspark.sql.functions.try_subtract(left: ColumnOrName, right: ColumnOrName) → pyspark.sql.column.Column[source]¶

Returns left-right and the result is null on overflow. The acceptable input types are the same with the - operator.

New in version 3.5.0.

Parameters

leftColumn or str
rightColumn or str

Examples

>>> df = spark.createDataFrame([(6000, 15), (1990, 2)], ["a", "b"])
>>> df.select(try_subtract(df.a, df.b).alias('r')).collect()
[Row(r=5985), Row(r=1988)]

>>> from pyspark.sql.types import StructType, StructField, IntegerType, StringType
>>> schema = StructType([
...     StructField("i", IntegerType(), True),
...     StructField("d", StringType(), True),
... ])
>>> df = spark.createDataFrame([(1, '2015-09-30')], schema)
>>> df = df.select(df.i, to_date(df.d).alias('d'))
>>> df.select(try_subtract(df.d, df.i).alias('r')).collect()
[Row(r=datetime.date(2015, 9, 29))]

>>> df.select(try_subtract(df.d, make_interval(df.i)).alias('r')).collect()
[Row(r=datetime.date(2014, 9, 30))]

>>> df.select(
...     try_subtract(df.d, make_interval(lit(0), lit(0), lit(0), df.i)).alias('r')
... ).collect()
[Row(r=datetime.date(2015, 9, 29))]

>>> df.select(
...     try_subtract(make_interval(df.i), make_interval(df.i)).alias('r')
... ).show(truncate=False)
+---------+
|r        |
+---------+
|0 seconds|
+---------+

pyspark.sql.functions.try_multiply

pyspark.sql.functions.try_sum