Question

我想将100k +行导入SQL Server表。

我有这样的插入（观察第6个子查询值）：

INSERT INTO BD_S3I.dbo.AGENDA 
(COD_UNDFBR, COD_DCPLNA, COD_TECNCA, COD_ATVIDE, DAT_PROGM_AGENDA, NUM_SQNCL_AGENDA, DAT_FINAL_AGENDA, COD_OCORR, COD_ROTA, NUM_SEMAN_PRGINS, NUM_DIAIN_PRGINS, DAT_INIC_PRGINS, MRC_SITUA_AGENDA, DAT_SUSPN_AGENDA, DAT_CONCL_AGENDA, DAT_REPRG_AGENDA, DCR_SITUA_AGENDA, DCR_AGENDA, MRC_AVISO_AGENDA, MRC_NEGLG_AGENDA, NUM_PERIO_PRGINS, DAT_DIAIN_PRGINS, DAT_JUSTN_AGENDA, COD_MTVNVS, MRC_ERP_AGENDA, COD_USUS3I_JUSTN)

VALUES

(1, 290, 2, 6, '2017-09-11 00:00:00.000', (SELECT CASE WHEN MAX(AGENDA.NUM_SQNCL_AGENDA) + 1 IS NULL THEN 1 ELSE MAX(AGENDA.NUM_SQNCL_AGENDA) + 1 END FROM AGENDA WHERE AGENDA.COD_UNDFBR = 1 AND AGENDA.COD_DCPLNA = 290 AND AGENDA.COD_TECNCA = 2 AND AGENDA.COD_ATVIDE = 6 AND AGENDA.DAT_PROGM_AGENDA = '2017-09-11 00:00:00.000'), '2017-09-17 00:00:00.000', NULL, 492, NULL, NULL, '2017-07-24 08:30:00.000', 'P', NULL, NULL, NULL, NULL, NULL, 'S', 'S', 7, '2017-07-24 00:00:00.000', NULL, NULL, 'N', NULL);

我将100k插入物放在彼此之下并开始导入。它工作正常，但是花费太多时间来执行所有100k +行。

我在考虑使用导入向导（时间更好？）。

问题是当我选择带有我的数据的excel文件时，导入向导不理解该值的子查询。它称之为长篇大论。

Answer 1

选择至少一个需要插入第6列的返回类型。就像

( SELECT x = CASE .... )

或者在最后使用返回类型和子查询。

Answer 2

只需将您的INSERT...VALUES转换为INSERT...SELECT即可，因为所有其他值都是标量，并且可以与子查询的SELECT语句一起包含在内：

INSERT INTO BD_S3I.dbo.AGENDA (COD_UNDFBR, COD_DCPLNA, COD_TECNCA, COD_ATVIDE, 
                               DAT_PROGM_AGENDA, NUM_SQNCL_AGENDA, DAT_FINAL_AGENDA, 
                               COD_OCORR, COD_ROTA, NUM_SEMAN_PRGINS, NUM_DIAIN_PRGINS, 
                               DAT_INIC_PRGINS, MRC_SITUA_AGENDA, DAT_SUSPN_AGENDA, 
                               DAT_CONCL_AGENDA, DAT_REPRG_AGENDA, DCR_SITUA_AGENDA, 
                               DCR_AGENDA, MRC_AVISO_AGENDA, MRC_NEGLG_AGENDA, 
                               NUM_PERIO_PRGINS, DAT_DIAIN_PRGINS, DAT_JUSTN_AGENDA, 
                               COD_MTVNVS, MRC_ERP_AGENDA, COD_USUS3I_JUSTN)

SELECT 1, 290, 2, 6, '2017-09-11 00:00:00.000', 
       CASE WHEN MAX(AGENDA.NUM_SQNCL_AGENDA) + 1 IS NULL 
            THEN 1 
            ELSE MAX(AGENDA.NUM_SQNCL_AGENDA) + 1 
       END,
       '2017-09-17 00:00:00.000', NULL, 492, NULL, 
        NULL, '2017-07-24 08:30:00.000', 'P', 
        NULL, NULL, NULL, NULL, NULL, 'S', 'S', 7, 
        '2017-07-24 00:00:00.000', NULL, NULL, 'N', NULL

FROM AGENDA 
WHERE AGENDA.COD_UNDFBR = 1 AND AGENDA.COD_DCPLNA = 290 
  AND AGENDA.COD_TECNCA = 2 AND AGENDA.COD_ATVIDE = 6 
  AND AGENDA.DAT_PROGM_AGENDA = '2017-09-11 00:00:00.000')

带子查询的导入向导

2 个答案: