1

我有以下数据框:

+-------------+-----------------+------------------+
|longitude    |latitude         |geom              |
+-------------+-----------------+------------------+
|-7.07378166  |33.826661        [00 00 00 00 01 0..|
|-7.5952683   |33.544191        [00 00 00 00 01 0..|                  
+-------------+-----------------+------------------+

我正在使用以下代码:

Dataset<Row> result_f = sparkSession.sql("select * from data_f where  ST_WITHIN(ST_GeomFromText(CONCAT('POINT(',longitude_f,' ',latitude_f,')',4326)),geom)");
result_f.show();

但我收到以下错误:

java.lang.ClassCastException: [B cannot be cast to org.apache.spark.sql.catalyst.util.ArrayData
at org.apache.spark.sql.geosparksql.expressions.ST_Within.eval(Predicates.scala:105)

编辑

longitude : Double type
latitude  : Double type
geom      : Binary type

任何想法 ?我需要你的帮助

谢谢

4

1 回答 1

2

我不认为 ST_GeomFromText 可用于从文本构造几何,但是有:

  • ST_GeomFromWKT
  • ST_GeomFromWKB
  • ST_GeomFromGeoJSON
  • ST_Point
  • ST_PointFromText
  • ST_PolygonFromText
  • ST_LineStringFromText
  • ST_PolygonFromEnvelope
  • ST_Circle

我建议使用ST_PointorST_PointFromText和之后的谓词ST_WITHIN

像这样的东西:

Dataset<Row> result_f = sparkSession.sql("select * from data_f where  ST_WITHIN(ST_Point(CAST(data_f.latitude AS Decimal(24,20)), CAST(data_f.longitude AS Decimal(24,20))),geom)");
result_f.show();
于 2020-07-23T06:43:03.750 回答