1

我正在尝试加入一个表和一个流并创建另一个表，如下所示：

CREATE TABLE table_fx_latest AS
   SELECT t1.currencyid,
          t1.maxtimestamp,
          t2.midprice 
  FROM stream_fx2 t2 LEFT JOIN table_fx_latest3 t1 
  ON t1.currencyid = t2.currencyid AND 
     t1.timestamp = t2.maxtimestamp 
  GROUP BY t1.currencyid, 
           t1.maxtimestamp, 
           t2.midprice;

但报如下错误：

Cannot RUN execution plan for this statement, CreateTableAsSelect{name=TABLE_FX_LATEST_PRICE6, query=Query{queryBody=QuerySpecification{select=Select{distinct=false, selectItems=[T1.CURRENCYID T1_CURRENCYID, T1.MAXTIMESTAMP MAXTIMESTAMP, T2.MIDPRICE MIDPRICE]}, from=Join{type=LEFT, left=AliasedRelation{relation=STREAM_FX2, alias=T2}, right=AliasedRelation{relation=TABLE_FX_LATEST3, alias=T1}, criteria=Optional[JoinOn{((T1.CURRENCYID = T2.CURRENCYID) AND (T2.TIMESTAMP = T1.MAXTIMESTAMP))}]}, =null, where=null, groupBy=Optional[GroupBy{isDistinct=false, groupingElements=[SimpleGroupBy{columns=[T1.CURRENCYID]}, SimpleGroupBy{columns=[T1.MAXTIMESTAMP]}, SimpleGroupBy{columns=[T2.MIDPRICE]}]}], having=null, orderBy=[], limit=null}, orderBy=[]}, notExists=false, properties={}}
Caused by: io.confluent.ksql.parser.tree.LogicalBinaryExpression cannot be cast to io.confluent.ksql.parser.tree.ComparisonExpression

这是stream_fx2流和table_fx_latest3表的描述：

ksql> describe stream_fx2;

Field      | Type
----------------------------------------
ROWTIME    | BIGINT           (system)
ROWKEY     | VARCHAR(STRING)  (system)
ID         | INTEGER
CURRENCY   | VARCHAR(STRING)
CURRENCYID | INTEGER
TIMESTAMP  | BIGINT
BIDPRICE   | DOUBLE
MIDPRICE   | DOUBLE
OFFERPRICE | DOUBLE

ksql> describe table_fx_latest3;

Field        | Type
------------------------------------------
ROWTIME      | BIGINT           (system)
ROWKEY       | VARCHAR(STRING)  (system)
CURRENCYID   | INTEGER          (key)
MAXTIMESTAMP | BIGINT
------------------------------------------

我猜这可能是 KSQL 的一个错误（仍在开发人员预览中），但我想确保我没有遗漏任何东西。任何帮助将非常感激。

4

1 回答 1

5

JOIN标准应该只是关键平等。如果您有任何其他标准，则需要将它们放在WHERE子句中。尝试以下操作：

CREATE TABLE table_fx_latest AS
  SELECT t1.currencyid,
         t1.maxtimestamp,
         t2.midprice 
  FROM stream_fx2 t2 LEFT JOIN table_fx_latest3 t1 
    ON t1.currencyid = t2.currencyid
  WHERE 
    t1.timestamp = t2.maxtimestamp 
  GROUP BY t1.currencyid, 
           t1.maxtimestamp, 
           t2.midprice;

让我知道这是否适合您。

于 2018-02-12T18:58:11.943 回答