使用 java 语言读取 spark 中的二进制列

Question

我有一个 DataFrame 女巫包含一个Binary列类型。

DataFrame：

+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|BinaryGeometry                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         
+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|[00 00 00 00 01 03 00 00 00 01 00 00 00 11 00 00 00 04 00 F0 00 DC CC 1A C0 87 14 01 81 1E 1B 41 40 FC FF EF 00 68 AA 1A C0 BF EE 57 20 85 19 41 40 04 00 F0 00 8C 86 1A C0 CC DC 8B DC AE 1A 41 40 FF FF EF 00 44 74 1A C0 CA 9D 5D 61 10 1C 41 40 FF FF EF 00 64 63 1A C0 BF 1F 98 0B 3A 1D 41 40 FF FF EF 00 44 47 1A C0 E4 6B A0 DD CE 1D 41 40 FC FF EF 00 D8 2B 1A C0 54 E4 71 67 6D 1C 41 40 FF FF EF 00 44 1A 1A C0 BF 1F 98 0B 3A 1D 41 40 02 00 F0 00 80 0B 1A C0 0D 80 00 13 2F 23 41 40 02 00 F0 00 B0 35 1A C0 CC F6 23 F8 BD 26 41 40 04 00 F0 00 0C 43 1A C0 73 1A 44 AF 16 26 41 40 02 00 F0 00 40 5A 1A C0 FF 54 9C 7C 2D 27 41 40 02 00 F0 00 50 68 1A C0 87 6E B9 42 44 28 41 40 02 00 F0 00 00 7C 1A C0 78 2B 85 BA F5 26 41 40 FC FF EF 00 18 91 1A C0 49 96 6F 58 C6 28 41 40 02 00 F0 00 B0 BC 1A C0 91 FA 4B 0E 7F 20 41 40 04 00 F0 00 DC CC 1A C0 87 14 01 81 1E 1B 41 40] |
+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+

我正在尝试阅读此专栏以提取几何格式。

根据我的研究，我发现 geospark 库中有一个 function，它是 ST_GeomFromWKB，它需要一个长二进制文件作为参数。

所以我正在执行以下代码：

df.withColumn("BinaryGeometry", hex(col("BinaryGeometry")))
                    .withColumn("BinaryGeometry",expr("ST_GeomFromWKB(BinaryGeometry)"))

我得到以下 output 女巫不正确：

POINT(0 0)

如何阅读本专栏以获得正确的几何值？

编辑

MySQL上的立柱结构

当我跟踪数据时，它看起来像这样：

当我点击[GEOMETRY - 113 o]时，会下载一个文本文件。

文本文件包含以下数据：

             ے FPہ>Pے{ح@@€àOہZDچ4ح@@‏ہ¨KہإضآTح@@‏àمJہî=Oƒح@@û ڑLہCBsn«ح@@ے FPہ>Pے{ح@@

编辑2

我在 MySQL 数据库上有这张表

+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|BinaryGeometry
+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|'POLYGON((-6.70005799736828 34.2118684058451,-6.66641236748546 34.1993751935347,-6.63139344658703 34.208461349766,-6.6135406633839 34.2192498881446,-6.5970611711964 34.2283339016717,-6.5695953508839 34.2328755410488,-6.54281617607921 34.2220887476075,-6.5256500383839 34.2283339016717,-6.51123048271984 34.2748740913813,-6.55242921318859 34.3026724029165,-6.56547547783703 34.2975672800575,-6.58813477959484 34.3060756457653,-6.60186768975109 34.314583149474,-6.62109376396984 34.3043740415805,-6.64169312920421 34.318553022848,-6.68426515068859 34.2538774367323,-6.70005799736828 34.2118684058451))',0
+-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

当我在 Spark 中加载此表时，我得到：

+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|BinaryGeometry                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         
+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|[00 00 00 00 01 03 00 00 00 01 00 00 00 11 00 00 00 04 00 F0 00 DC CC 1A C0 87 14 01 81 1E 1B 41 40 FC FF EF 00 68 AA 1A C0 BF EE 57 20 85 19 41 40 04 00 F0 00 8C 86 1A C0 CC DC 8B DC AE 1A 41 40 FF FF EF 00 44 74 1A C0 CA 9D 5D 61 10 1C 41 40 FF FF EF 00 64 63 1A C0 BF 1F 98 0B 3A 1D 41 40 FF FF EF 00 44 47 1A C0 E4 6B A0 DD CE 1D 41 40 FC FF EF 00 D8 2B 1A C0 54 E4 71 67 6D 1C 41 40 FF FF EF 00 44 1A 1A C0 BF 1F 98 0B 3A 1D 41 40 02 00 F0 00 80 0B 1A C0 0D 80 00 13 2F 23 41 40 02 00 F0 00 B0 35 1A C0 CC F6 23 F8 BD 26 41 40 04 00 F0 00 0C 43 1A C0 73 1A 44 AF 16 26 41 40 02 00 F0 00 40 5A 1A C0 FF 54 9C 7C 2D 27 41 40 02 00 F0 00 50 68 1A C0 87 6E B9 42 44 28 41 40 02 00 F0 00 00 7C 1A C0 78 2B 85 BA F5 26 41 40 FC FF EF 00 18 91 1A C0 49 96 6F 58 C6 28 41 40 02 00 F0 00 B0 BC 1A C0 91 FA 4B 0E 7F 20 41 40 04 00 F0 00 DC CC 1A C0 87 14 01 81 1E 1B 41 40] |
+----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+

然后获取原始数据'Polygon............'我执行以下代码：

df.withColumn("geom",expr("ST_GeomFromWKB(BinaryGeometry)"));

但我收到以下错误：

20/08/10 22:28:50 ERROR Executor: Exception in task 87.0 in stage 39.0 (TID 929) java.lang.ClassCastException: [B cannot be cast to org.apache.spark.unsafe.types.UTF8String
at org.apache.spark.sql.geosparksql.expressions.ST_GeomFromWKB.eval(Constructors.scala:174)
at org.apache.spark.sql.catalyst.expressions.GeneratedClass$SpecificUnsafeProjection.writeFields_0_39$(Unknown Source)
at org.apache.spark.sql.catalyst.expressions.GeneratedClass$SpecificUnsafeProjection.apply(Unknown Source)
at org.apache.spark.sql.catalyst.expressions.GeneratedClass$SpecificUnsafeProjection.apply(Unknown Source)

编辑3

Answer 1

MySQL 几何类型并未真正存储为WKB ，因此无法使用 geospark 的方法ST_GeomFromWKB 。

https://dev.mysql.com/doc/refman/5.7/en/gis-data-formats.html

在内部，MySQL 以不同于WKT或WKB格式的格式存储几何值。 （内部格式类似于WKB ，但有一个初始的 4 个字节来指示 SRID。）

解决方案是在加载 dataframe 时指定查询，该查询将使用内置 function ST_AsWKT将geometry类型解析为WKT格式：

df = spark
  .read
  .format("jdbc")
  .options(
    Map(
      "driver" -> "com.mysql.cj.jdbc.Driver",
      "url" -> "jdbc:mysql://host:3306/db",
      "user" -> "user",
      "password" -> "password",
      "dbtable" -> "(select ST_AsWKT(BinaryGeometry) as BinaryGeometry from geo_table) as t")
  )
  .load
  .withColumn("BinaryGeometry",expr("ST_GeomFromWKT(BinaryGeometry)"))

或者，可以使用 WKB 将列提取为

select hex(ST_AsWKB(BinaryGeometry)) as BinaryGeometry from geo_table

并馈送到ST_GeomFromWKB方法。

scala> df.printShema()
root
 |-- BinaryGeometry: geometry (nullable = false)

scala> df.show()
+--------------------+
|      BinaryGeometry|
+--------------------+
|POLYGON ((-7.5783...|
+--------------------+