创建sessionWindowEngine报错‘表不能含重复列名’的排查与解决

创建sessionWindowEngine时服务器报错：A table can't contain duplicate column names，但输入表与输出表均无重复列名，求报错原因及脚本修改方案。

输入表定义 #

ofpanel_colnames = [
"TimeStamp_1",
'ID_1', 'UpperLimit_1', 'LowerLimit_1',
'PreOpenInt_1',
'OpenPrice_1', 'HighPrice_1', 'LowPrice_1', 'LastPrice_1',
'BidPrice_1', 'BidVol_1', 'AskPrice_1', 'AskVol_1',
'Volume_1', 'Amount_1', 'OpenInt_1',
'ClosePrice_1', 'SettlePrice_1',
'ReceiveTime_1',
'UnderlyingID',
"TimeStamp_2",
'ID_2', 'UpperLimit_2', 'LowerLimit_2',
'BidPrice_2', 'BidVol_2', 'AskPrice_2', 'AskVol_2',
'ReceiveTime_2'
]
ofpanel_colTypes = [
TIMESTAMP,
SYMBOL, DOUBLE, DOUBLE,
DOUBLE,
DOUBLE, DOUBLE, DOUBLE, DOUBLE,
DOUBLE, INT, DOUBLE, INT,
INT, DOUBLE, DOUBLE,
DOUBLE, DOUBLE,
NANOTIMESTAMP,
STRING,
TIMESTAMP,
SYMBOL, DOUBLE, DOUBLE,
DOUBLE, INT, DOUBLE, INT,
NANOTIMESTAMP
]
share streamTable(1:0, ofpanel_colnames, ofpanel_colTypes ) as ofPanel_stream

输出表定义 #

try{dropStreamTable("SampleStream")}catch(ex){print(ex)}
share streamTable(
    1:0,
    ["LogTime","InstrumentID","TimeStamp_1","ID_1","UpperLimit_1","LowerLimit_1","PreOpenInt_1"],
    [TIMESTAMP,STRING,TIMESTAMP,STRING,DOUBLE, DOUBLE, INT]
) as SampleStream

引擎创建脚本 #

try{dropStreamEngine(name="DataProcessor")}catch(ex){print(ex)}
engine_ = createSessionWindowEngine(
    name = 'DataProcessor', 
    sessionGap = 30*1000,  // 30s response window
    metrics=<[
        last(TimeStamp_1),
        last(ID_1),
        last(UpperLimit_1),
        last(LowerLimit_1),
        last(PreOpenInt_1)
    ]>, 
    dummyTable = ofPanel_stream, 
    outputTable = SampleStream,  
    useSystemTime=true, 
    keyColumn = `ID_1,
    useSessionStartTime=true
)

报错原因 #

1. key列重复输出
   指定keyColumn = ID_1后，引擎会自动将ID_1作为输出结果的第一列；但你在metrics中又添加了last(ID_1)，这会在metrics结果中再生成一个ID_1列，最终输出结果中出现两个ID_1`列，触发重复列名错误。

2. 列顺序不匹配
   当useSessionStartTime=true时，引擎默认会添加名为LogTime的会话开始时间列，紧随key列之后。你的输出表列顺序是LogTime在前、ID_1在后，与引擎实际输出顺序（ID_1 → LogTime → metrics列）完全不符，进一步加剧了列名冲突。

3. 未匹配的自定义列
   输出表中的InstrumentID列在metrics中没有对应计算逻辑，引擎无法自动填充该列，后续也会引发数据写入错误。

脚本修改方案 #

步骤1：调整metrics，移除重复的key列计算 #

删除metrics中的last(ID_1)，因为key列会由引擎自动输出。

步骤2：调整输出表列顺序，匹配引擎输出逻辑 #

引擎输出顺序为：key列（ID_1） → 会话开始时间列（LogTime） → metrics列 → 自定义列（需在metrics中添加对应计算）。这里假设InstrumentID对应输入表的UnderlyingID，在metrics中添加last(UnderlyingID)来填充。

修改后的完整脚本 #

1. 重新定义输出表

try{dropStreamTable("SampleStream")}catch(ex){print(ex)}
share streamTable(
    1:0,
    ["ID_1","LogTime","TimeStamp_1","UpperLimit_1","LowerLimit_1","PreOpenInt_1","InstrumentID"],
    [STRING,TIMESTAMP,TIMESTAMP,DOUBLE, DOUBLE, INT,STRING]
) as SampleStream

2. 重新创建sessionWindowEngine

try{dropStreamEngine(name="DataProcessor")}catch(ex){print(ex)}
engine_ = createSessionWindowEngine(
    name = 'DataProcessor', 
    sessionGap = 30*1000,  // 30s response window
    metrics=<[
        last(TimeStamp_1),
        last(UpperLimit_1),
        last(LowerLimit_1),
        last(PreOpenInt_1),
        last(UnderlyingID)  // 对应输出表的InstrumentID列
    ]>, 
    dummyTable = ofPanel_stream, 
    outputTable = SampleStream,  
    useSystemTime=true, 
    keyColumn = `ID_1,
    useSessionStartTime=true
)

可选方案：自定义会话开始时间列名 #

如果希望保留原输出表的列顺序，可以指定sessionStartTimeColumn参数自定义会话开始时间列名，同时调整metrics和输出表列对应关系：

输出表修改

try{dropStreamTable("SampleStream")}catch(ex){print(ex)}
share streamTable(
    1:0,
    ["LogTime","InstrumentID","ID_1","TimeStamp_1","UpperLimit_1","LowerLimit_1","PreOpenInt_1"],
    [TIMESTAMP,STRING,STRING,TIMESTAMP,DOUBLE, DOUBLE, INT]
) as SampleStream

引擎创建修改

try{dropStreamEngine(name="DataProcessor")}catch(ex){print(ex)}
engine_ = createSessionWindowEngine(
    name = 'DataProcessor', 
    sessionGap = 30*1000,  // 30s response window
    metrics=<[
        last(ID_1),
        last(TimeStamp_1),
        last(UpperLimit_1),
        last(LowerLimit_1),
        last(PreOpenInt_1),
        last(UnderlyingID)
    ]>, 
    dummyTable = ofPanel_stream, 
    outputTable = SampleStream,  
    useSystemTime=true, 
    keyColumn = `ID_1,
    useSessionStartTime=true,
    sessionStartTimeColumn = `LogTime  // 指定会话开始时间列名与输出表匹配
)

内容的提问来源于stack exchange，提问作者user31077149