Python 如何通过对前一行和当前行的值求和来填充 PySpark Dataframe 的行

你可以取coalesce(level, change)英镑的总和.

import pyspark.sql.functions as func
from pyspark.sql.window import Window as wd

data_sdf. \
    withColumn('new_stock_level', 
               func.sum(func.coalesce('stock_level', 'stock_change')).
               over(wd.partitionBy('cat').orderBy('time').rowsBetween(-sys.maxsize, 0))
               ). \
    show()

# +------+----+-----------+------------+---------------+
# |   cat|time|stock_level|stock_change|new_stock_level|
# +------+----+-----------+------------+---------------+
# | apple|   1|          4|        null|              4|
# | apple|   2|       null|          -2|              2|
# | apple|   3|       null|           5|              7|
# |banana|   1|         12|        null|             12|
# |banana|   2|       null|           4|             16|
# |orange|   1|          1|        null|              1|
# |orange|   2|       null|          -7|             -6|
# +------+----+-----------+------------+---------------+