SparkSQL执行SQL触发ParseException：别名单引号问题求助

错误原因 #

Spark SQL里，单引号'仅用来表示字符串常量，不能定义列别名。你代码里用AS 'PhaseID'这种写法，会被SQL解析器当成字符串值，而非列的别名，直接触发语法解析错误。

解决方案 #

把列别名的单引号替换为反引号（`），无空格的别名也可以直接省略引号：

• 无空格别名（如PhaseID）：AS PhaseID
• 带空格别名（如Phase Assignee）：AS Phase Assignee``

修改后的代码 #

query = f"""
SELECT ROW_NUMBER() OVER (ORDER BY  (SELECT 1)) + {pDimensionkey} AS ibfPhaseAssignmentKey, Base.* FROM  (
SELECT DISTINCT
Phase.lxOid AS PhaseID
, TAName.mxFullname AS `Phase Assignee`
, DurationAttr.lxVal AS `Phase Estimated Duration`
, EstStartDate.lxVal AS `Phase Estimated Start Date`
, EstFinishDate.lxVal AS `Phase Estimated Finish Date`
, ActStartDate.lxVal AS `Phase Actual Start Date`
, ActEndDate.lxVal AS `Phase Actual Completion Date`
, relatedTask.lxOid AS `Related Object ID`
, pr.ProjectId AS `Related Project ID`
FROM lippertus_Enoviasilver.lxBO_49243908 Phase
LEFT JOIN lippertus_Enoviasilver.mxBusType PhaseType ON PhaseType.mxOid = Phase.lxType AND Phase.lxType = '1071955709'
LEFT JOIN lippertus_Enoviasilver.lxRO_49243908 TARO ON Phase.lxOid = TARO.lxToId AND TARO.lxType = '1968973284'
LEFT JOIN lippertus_Enoviasilver.lxBO_49243908 TA ON TARO.lxFromId = TA.lxOid
LEFT JOIN lippertus_Enoviasilver.mxUser TAName ON TA.lxName = TAName.mxName
LEFT JOIN lippertus_Enoviasilver.mxBusType TAType ON TA.lxType = TAType.mxOid
LEFT JOIN lippertus_Enoviasilver.lxReal_49243908 DurationAttr ON Phase.lxOid = DurationAttr.lxOid AND DurationAttr.lxType = '-668419160'
LEFT JOIN lippertus_Enoviasilver.lxDate_49243908 EstStartDate ON Phase.lxOid = EstStartDate.lxOid AND EstStartDate.lxType = '-668414286'
LEFT JOIN lippertus_Enoviasilver.lxDate_49243908 EstFinishDate ON Phase.lxOid = EstFinishDate.lxOid AND EstFinishDate.lxType = '-668416845'
LEFT JOIN lippertus_Enoviasilver.lxDate_49243908 ActStartDate ON Phase.lxOid = ActStartDate.lxOid AND ActStartDate.lxType = '-308477270'
LEFT JOIN lippertus_Enoviasilver.lxDate_49243908 ActEndDate ON Phase.lxOid = ActEndDate.lxOid AND ActEndDate.lxType = '1454168401'
LEFT JOIN lippertus_Enoviasilver.lxRO_49243908 PhaseRel ON Phase.lxOid = PhaseRel.lxFromId
LEFT JOIN lippertus_Enoviasilver.lxBO_49243908 relatedTask ON PhaseRel.lxToId = relatedTask.lxOid
LEFT JOIN lippertus_Enoviasilver.lxRO_49243908 phasesRO ON phasesRO.lxToId = phase.lxOid
LEFT JOIN lippertus_Enoviasilver.mxRelType projectRelType ON projectRelType.mxOid = phasesRO.lxType AND projectRelType.mxName = 'Subtask'
LEFT JOIN empower.Project pr ON phasesRO.lxFromId = pr.ProjectId) Base"""
source_query = spark.sql(query)

补充说明 #

你之前尝试的'''是Python的多字符串包裹语法，和SQL里的别名规则无关，所以解决不了问题。Spark SQL对标识符（别名、表名等）的规则是：

• 无特殊字符（空格、关键字）的标识符可以直接书写
• 带特殊字符或关键字的标识符必须用反引号包裹（兼容Hive语法）

内容的提问来源于stack exchange，提问作者Shuo Xu