From f66fe6b2dfe34190fd7bdc31ecbdd37f1be08730 Mon Sep 17 00:00:00 2001
From: WilliamZhu <allwefantasy@gmail.com>
Date: Thu, 23 Dec 2021 14:48:54 +0800
Subject: [PATCH] convert options in include statment to set statments (#1633)

* convert options in include statment to set statments

* add test and fix exception throws when no suffix in  include statement

* ignore result check in pr_1633.mlsql

revise the code example

add checking multi group params

revise the remind msg when setting multi-group params in the Discretizer et

rename the checking functions
---
 .../dsl/mmlib/algs/SQLDiscretizer.scala       | 21 +++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)

diff --git a/streamingpro-mlsql/src/main/java/streaming/dsl/mmlib/algs/SQLDiscretizer.scala b/streamingpro-mlsql/src/main/java/streaming/dsl/mmlib/algs/SQLDiscretizer.scala
index 54d9f9d3e..f8fb95f8d 100644
--- a/streamingpro-mlsql/src/main/java/streaming/dsl/mmlib/algs/SQLDiscretizer.scala
+++ b/streamingpro-mlsql/src/main/java/streaming/dsl/mmlib/algs/SQLDiscretizer.scala
@@ -87,8 +87,17 @@ class SQLDiscretizer(override val uid: String) extends SQLAlg with Functions wit
     return metas
   }
 
+  def checkWithoutGroupParams(params: Map[String, String]): Unit = {
+    val group = params.map(p => {
+      val key = p._1
+      key.contains(DiscretizerParamsConstrant.PARAMS_PREFIX + ".")
+    }).filter(_ == true)
+    require(group.size == 0, "The multi-group params are not available for the Discretizer Et!")
+  }
+
   def trainWithoutGroup(df: DataFrame, params: Map[String, String], _method: String, metaPath: String): DataFrame = {
     val dfWithId = df.withColumn("id", monotonically_increasing_id)
+    checkWithoutGroupParams(params)
     var transformedDF = dfWithId
     // we need save metadatas with index, because we need index
     val metas: Array[(Int, DiscretizerTrainData)] = {
@@ -220,10 +229,14 @@ class SQLDiscretizer(override val uid: String) extends SQLAlg with Functions wit
       |
       |train data1 as Discretizer.`/tmp/model`
       |where method="bucketizer"
-      |and `fitParam.0.inputCol`="a"
-      |and `fitParam.0.splitArray`="-inf,0.0,1.0,inf"
-      |and `fitParam.1.inputCol`="b"
-      |and `fitParam.1.splitArray`="-inf,0.0,1.0,inf";
+      |and `inputCol`="a"
+      |and `splits`="-inf,0.0,1.0,inf";
+      |
+      |train data1 as Discretizer.`/tmp/discretizer`
+      |where method="quantile"
+      |and `inputCol`="a"
+      |and `outputCol`="a_v"
+      |and `numBuckets`="3";
       |;
     """.stripMargin)