Python pyspark columns count

Question

我有一个这样的数据集:

区2

将数据帧堆叠，然后执行groupby+count

expr = "stack(2, 'Zone 1', `Zone 1`, 'Zone 2', `Zone 2`) as (zone, category)"
result = df.selectExpr(expr).groupBy('category', 'zone').count()

结果

+--------+------+-----+
|category|  zone|count|
+--------+------+-----+
|       A|Zone 1|    3|
|       A|Zone 2|    2|
|       B|Zone 2|    3|
|       B|Zone 1|    2|
+--------+------+-----+