2

我是 spark 的初学者，我陷入了如何使用数据框发出 sql 请求。

我有以下两个数据框。

dataframe_1
+-----------------+-----------------+----------------------+---------------------+
|id               |geometry_tyme    |geometry              |rayon                |
+-----------------+-----------------+----------------------+---------------------+
|50               |Polygon          |[00 00 00 00 01 0...] |200                  |
|54               |Point            |[00 00 00 00 01 0.. ] |320179               |
+-----------------+-----------------+----------------------+---------------------+
dataframe_2
+-----------------+-----------------+----------------------+
|id2              |long             |lat                   |               
+-----------------+-----------------+----------------------+
|[70,50,600,]     | -9.6198783      |44.5942549            |
|[20,140,39,]     |-6.6198783       |44.5942549            |
+-----------------+-----------------+----------------------+

我想执行以下请求。

"SELECT dataframe_1.* FROM dataframe_1 WHERE dataframe_1.id IN ("
                            + id2
                            + ") AND ((dataframe_1.geometry_tyme='Polygon' AND (ST_WITHIN(ST_GeomFromText(CONCAT('POINT(',"
                            + long
                            + ",' ',"
                            + lat
                            + ",')'),4326),dataframe_1.geometry))) OR ( (dataframe_1.geometry_tyme='LineString' OR dataframe_1.geomType='Point') AND     ST_Intersects(ST_buffer(dataframe_1.geom,(dataframe_1.rayon/100000)),ST_GeomFromText(CONCAT('POINT(',"
                            + long
                            + ",' ',"
                            + lat
                            + ",')'),4326)))) "

我真的卡住了，我应该加入两个数据框还是什么？我尝试像这样使用 id 和 idZone 加入两个数据框：

     dataframe_2.select(explode(col("id2").as ("id2"))).join(dataframe_1,col("id2").equalTo(dataframe_1.col("id")));

但在我看来，加入并不是正确的选择。

我需要你帮忙 。

谢谢

4

1 回答 1

1

1.从您的数据框创建临时视图。

dataframe_1.createOrReplaceTempView("dataframe_1")
dataframe_2.createOrReplaceTempViews("dataframe_2")

2.运行你的SQLfinal_df = spark.sql("your SQL here")

于 2020-07-13T20:36:11.550 回答