我们有一个查询，其中包含针对不同日期时间列的15个convert函数调用，以及其他20个列。使用Convert函数执行查询时需要将近5分钟，但是当我们注释掉/删除转换函数时，它只需要15秒。
CONVERT (VARCHAR (10), TABLE.[COLUMN NAME], 101)
SELECT
T1.CONFLICT_ID,
CONVERT (VARCHAR (10),T1.LAST_UPD, 101) + ' ' + CONVERT (VARCHAR (10),T1.LAST_UPD, 8),
CONVERT (VARCHAR (10),T1.CREATED, 101) + ' ' + CONVERT (VARCHAR (10),T1.CREATED, 8),
T1.LAST_UPD_BY,
T1.CREATED_BY,
T1.MODIFICATION_NUM,
T1.ROW_ID,
CONVERT (VARCHAR (10),T1.CREATED, 101) + ' ' + CONVERT (VARCHAR (10),T1.CREATED, 8),
T1.ROW_ID,
CONVERT (VARCHAR (10),T1.LAST_UPD, 101) + ' ' + CONVERT (VARCHAR (10),T1.LAST_UPD, 8),
T1.LAST_UPD_BY,
CONVERT (VARCHAR (10),T1.TODO_AFTER_DT, 101) + ' ' + CONVERT (VARCHAR (10),T1.TODO_AFTER_DT, 8),
T1.APPT_REPT_TYPE,
T27.ATTRIB_47,
T1.ROW_STATUS,
T24.OWN_INST_ID,
T24.INTEGRATION_ID,
T15.SR_TITLE,
T27.ATTRIB_45,
T1.APPT_ALARM_TM_MIN,
T15.SR_NUM,
T1.PR_PRDINT_ID,
T1.PR_SR_ID,
T1.PR_SYMPTOM_CD,
CONVERT (VARCHAR (10),T13.PLAN_END_DT, 101) + ' ' + CONVERT (VARCHAR (10),T13.PLAN_END_DT, 8),
CONVERT (VARCHAR (10),T13.PLAN_START_DT, 101) + ' ' + CONVERT (VARCHAR (10),T13.PLAN_START_DT, 8),
T12.BL_CLASS_CD,
T12.BL_CURCY_CD,
T1.PROJ_ID,
CONVERT (VARCHAR (10),T13.ACTL_END_DT, 101) + ' ' + CONVERT (VARCHAR (10),T13.ACTL_END_DT, 8),
CONVERT (VARCHAR (10),T13.ACTL_START_DT, 101) + ' ' + CONVERT (VARCHAR (10),T13.ACTL_START_DT, 8),
T9.NAME,
T9.CURCY_CD,
T1.OPTY_ID,
T10.BASE_CURCY_CD,
T1.TARGET_OU_ID,
T10.LOC,
T10.NAME,
CONVERT (VARCHAR (10),T1.TODO_ACTL_END_DT, 101) + ' ' + CONVERT (VARCHAR (10),T1.TODO_ACTL_END_DT, 8),
T1.SRV_REGN_ID,
T16.NAME,
T18.AMS_ROUTE_TO_CD,
T1.ME_EVT_ID,
T1.ASGN_MANL_FLG,
T20.MSTR_CASE_ID,
T22.NUM_PAGE,
T18.AMS_ACT_ID,
T19.MAX_CAPABILITY_LVL,
T19.MIN_CAPABILITY_LVL,
T19.RTE_APPR_FLG,
T1.OWNER_LOGIN,
T2.BU_ID,
T1.OWNER_PER_ID,
T1.APPT_REPT_FLG,
CONVERT (VARCHAR (10),T1.APPT_REPT_END_DT, 101) + ' ' + CONVERT (VARCHAR (10),T1.APPT_REPT_END_DT, 8),
T1.PR_ORDER_ID,
T1.DO_NOT_ROUTE_FLG,
T1.TARGET_OU_ADDR_ID,
T1.ARREST_ID,
T1.ASSET_ID,
T1.ASGN_USR_EXCLD_FLG,
T7.PR_POSTN_ID,
T28.PIM_APPT_FLG,
T28.SEBL_APPT_FLG,
T28.UNSPRTD_REPT_FLG,
T26.SRM_REQUEST_ID,
T12.NAME,
T12.PROJ_NUM,
T1.X_ENTITY_TYPE,
T1.XTRNL_ACCESS_FLG,
T29.EVIDENCE_ID,
T1.SRA_TYPE_CD,
T1.COMM_ID,
T1.COMMENTS_LONG,
T1.COMMENTS,
T10.PR_ADDR_ID,
T1.PR_PROD_CAT_ID,
T3.FST_NAME,
T1.TARGET_PER_ID,
T3.LAST_NAME,
T1.OWNER_OU_ID,
T1.COMM_ID,
T1.CMPLTD_FLG,
T20.CASE_NUM,
T1.X_OFFENSE_ID,
T27.ATTRIB_37,
T1.PREV_ACT_ID,
CONVERT (VARCHAR (10),T1.APPT_END_DT, 101) + ' ' + CONVERT (VARCHAR (10),T1.APPT_END_DT, 8),
CONVERT (VARCHAR (10),T1.APPT_END_TM, 101) + ' ' + CONVERT (VARCHAR (10),T1.APPT_END_TM, 8),
T14.SCHED_ENGINE_CD,
T6.ASSET_NUM,
T11.NAME,
T1.OWNER_POSTN_ID,
T25.WEB_ACCESS_FLG,
T1.SRC_ID,
T1.X_CASE_ID,
T20.NAME,
CONVERT (VARCHAR (10),T1.TODO_DUE_DT, 101) + ' ' + CONVERT (VARCHAR (10),T1.TODO_DUE_DT, 8),
T1.APPT_DURATION_MIN,
CONVERT (VARCHAR (10),T18.TLR_EFFECTIVE_DT, 101) + ' ' + CONVERT (VARCHAR (10),T18.TLR_EFFECTIVE_DT, 8),
T26.ACTIVE_FLG,
T19.NAME,
T1.SRA_DEFECT_ID,
T1.SRA_SR_ID,
T1.TMPL_PLANITEM_ID,
T1.ACTIVITY_UID,
T1.AGREEMENT_ID,
T1.PYMNT_FLG,
T1.ALARM_FLAG,
T1.CREATOR_LOGIN,
T1.ASGN_DNRM_FLG,
T1.NAME,
T18.TLR_INTG_RET_CD,
T1.INCIDENT_ID,
T8.X_SESSION_ID,
T18.TLR_INTG_ID,
T18.TLR_INTG_MSG,
CONVERT (VARCHAR (10),T1.TODO_ACTL_END_DT, 101) + ' ' + CONVERT (VARCHAR (10),T1.TODO_ACTL_END_DT, 8),
T1.DONE_FLG,
CONVERT (VARCHAR (10),T1.APPT_START_DT, 101) + ' ' + CONVERT (VARCHAR (10),T1.APPT_START_DT, 8),
T4.POSTN_TYPE_CD,
CONVERT (VARCHAR (10),T1.STATUS_UPD_TS, 101) + ' ' + CONVERT (VARCHAR (10),T1.STATUS_UPD_TS, 8),
T10.DIVISION,
T1.COMMENTS,
T26.STAGE_ID,
CONVERT (VARCHAR (10),T1.APPT_START_TM, 101) + ' ' + CONVERT (VARCHAR (10),T1.APPT_START_TM, 8),
CONVERT (VARCHAR (10),T1.TODO_ACTL_START_DT, 101) + ' ' + CONVERT (VARCHAR (10),T1.TODO_ACTL_START_DT, 8),
T1.ASSESS_7,
T1.X_AUTO_FLG,
T1.X_SEQUENCE,
T23.FORMAT_TYPE_CD,
T1.X_OFFENSE_ID,
T5.EMP_ID,
T21.SRV_REGN_ID,
T2.FST_NAME,
T2.LAST_NAME,
T14.SCHED_PARAM_SET_ID,
T1.PAR_EVT_ID,
T1.BILLABLE_FLG,
T1.APPT_REPT_APPT_ID,
T1.APPT_REPT_REPL_CD,
T1.TEMPLATE_FLG,
T1.ASGN_SYS_FLG,
T1.PCT_COMPLETE,
T1.TARGET_PER_ADDR_ID,
T18.SUB_TYPE_CD,
T1.CL_SUBJ_ID,
T1.CAL_DISP_FLG,
CONVERT (VARCHAR (10),T1.TODO_PLAN_START_DT, 101) + ' ' + CONVERT (VARCHAR (10),T1.TODO_PLAN_START_DT, 8),
CONVERT (VARCHAR (10),T1.TODO_PLAN_END_DT, 101) + ' ' + CONVERT (VARCHAR (10),T1.TODO_PLAN_END_DT, 8),
T1.COST_CURCY_CD,
T15.SR_STAT_ID,
T1.SUBTYPE_CD,
T1.TODO_CD,
T20.X_CASE_MODE,
T1.CAL_TYPE_CD,
T1.EVT_STAT_CD,
T30.ROW_ID,
T31.ROW_STATUS,
T33.AMS_CAPABILITY_LVL,
T34.LOGIN,
T32.ROW_ID,
T36.GEOCD_VALID_FLG,
T22.ROW_ID,
T22.PAR_ROW_ID,
T22.MODIFICATION_NUM,
T22.CREATED_BY,
T22.LAST_UPD_BY,
CONVERT (VARCHAR (10),T22.CREATED, 101) + ' ' + CONVERT (VARCHAR (10),T22.CREATED, 8),
CONVERT (VARCHAR (10),T22.LAST_UPD, 101) + ' ' + CONVERT (VARCHAR (10),T22.LAST_UPD, 8),
T22.CONFLICT_ID,
T22.PAR_ROW_ID,
T18.ROW_ID,
T18.PAR_ROW_ID,
T18.MODIFICATION_NUM,
T18.CREATED_BY,
T18.LAST_UPD_BY,
CONVERT (VARCHAR (10),T18.CREATED, 101) + ' ' + CONVERT (VARCHAR (10),T18.CREATED, 8),
CONVERT (VARCHAR (10),T18.LAST_UPD, 101) + ' ' + CONVERT (VARCHAR (10),T18.LAST_UPD, 8),
T18.CONFLICT_ID,
T18.PAR_ROW_ID,
T23.ROW_ID,
T23.PAR_ROW_ID,
T23.MODIFICATION_NUM,
T23.CREATED_BY,
T23.LAST_UPD_BY,
CONVERT (VARCHAR (10),T23.CREATED, 101) + ' ' + CONVERT (VARCHAR (10),T23.CREATED, 8),
CONVERT (VARCHAR (10),T23.LAST_UPD, 101) + ' ' + CONVERT (VARCHAR (10),T23.LAST_UPD, 8),
T23.CONFLICT_ID,
T23.PAR_ROW_ID,
T24.ROW_ID,
T24.PAR_ROW_ID,
T24.MODIFICATION_NUM,
T24.CREATED_BY,
T24.LAST_UPD_BY,
CONVERT (VARCHAR (10),T24.CREATED, 101) + ' ' + CONVERT (VARCHAR (10),T24.CREATED, 8),
CONVERT (VARCHAR (10),T24.LAST_UPD, 101) + ' ' + CONVERT (VARCHAR (10),T24.LAST_UPD, 8),
T24.CONFLICT_ID,
T24.PAR_ROW_ID,
T25.ROW_ID,
T25.PAR_ROW_ID,
T25.MODIFICATION_NUM,
T25.CREATED_BY,
T25.LAST_UPD_BY,
CONVERT (VARCHAR (10),T25.CREATED, 101) + ' ' + CONVERT (VARCHAR (10),T25.CREATED, 8),
CONVERT (VARCHAR (10),T25.LAST_UPD, 101) + ' ' + CONVERT (VARCHAR (10),T25.LAST_UPD, 8),
T25.CONFLICT_ID,
T25.PAR_ROW_ID,
T26.ROW_ID,
T26.PAR_ROW_ID,
T26.MODIFICATION_NUM,
T26.CREATED_BY,
T26.LAST_UPD_BY,
CONVERT (VARCHAR (10),T26.CREATED, 101) + ' ' + CONVERT (VARCHAR (10),T26.CREATED, 8),
CONVERT (VARCHAR (10),T26.LAST_UPD, 101) + ' ' + CONVERT (VARCHAR (10),T26.LAST_UPD, 8),
T26.CONFLICT_ID,
T26.PAR_ROW_ID,
T27.ROW_ID,
T27.PAR_ROW_ID,
T27.MODIFICATION_NUM,
T27.CREATED_BY,
T27.LAST_UPD_BY,
CONVERT (VARCHAR (10),T27.CREATED, 101) + ' ' + CONVERT (VARCHAR (10),T27.CREATED, 8),
CONVERT (VARCHAR (10),T27.LAST_UPD, 101) + ' ' + CONVERT (VARCHAR (10),T27.LAST_UPD, 8),
T27.CONFLICT_ID,
T27.PAR_ROW_ID,
T28.ROW_ID,
T28.PAR_ROW_ID,
T28.MODIFICATION_NUM,
T28.CREATED_BY,
T28.LAST_UPD_BY,
CONVERT (VARCHAR (10),T28.CREATED, 101) + ' ' + CONVERT (VARCHAR (10),T28.CREATED, 8),
CONVERT (VARCHAR (10),T28.LAST_UPD, 101) + ' ' + CONVERT (VARCHAR (10),T28.LAST_UPD, 8),
T28.CONFLICT_ID,
T28.PAR_ROW_ID,
T29.ROW_ID,
T29.PAR_ROW_ID,
T29.MODIFICATION_NUM,
T29.CREATED_BY,
T29.LAST_UPD_BY,
CONVERT (VARCHAR (10),T29.CREATED, 101) + ' ' + CONVERT (VARCHAR (10),T29.CREATED, 8),
CONVERT (VARCHAR (10),T29.LAST_UPD, 101) + ' ' + CONVERT (VARCHAR (10),T29.LAST_UPD, 8),
T29.CONFLICT_ID,
T29.PAR_ROW_ID,
T30.ROW_ID,
T31.ROW_ID,
T32.ROW_ID,
T35.ROW_ID,
T36.ROW_ID
FROM
dbo.S_EVT_ACT T1
LEFT OUTER JOIN dbo.S_CONTACT T2 ON T1.OWNER_PER_ID = T2.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_CONTACT T3 ON T1.TARGET_PER_ID = T3.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_POSTN T4 ON T2.PR_HELD_POSTN_ID = T4.ROW_ID
LEFT OUTER JOIN dbo.S_ACT_EMP T5 ON T1.ROW_ID = T5.ACTIVITY_ID
LEFT OUTER JOIN dbo.S_ASSET T6 ON T1.ASSET_ID = T6.ROW_ID
LEFT OUTER JOIN dbo.S_CONTACT T7 ON T1.TARGET_PER_ID = T7.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_INCIDENT T8 ON T1.INCIDENT_ID = T8.ROW_ID
LEFT OUTER JOIN dbo.S_OPTY T9 ON T1.OPTY_ID = T9.ROW_ID
LEFT OUTER JOIN dbo.S_ORG_EXT T10 ON T1.TARGET_OU_ID = T10.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_POSTN T11 ON T1.OWNER_POSTN_ID = T11.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_PROJ T12 ON T1.PROJ_ID = T12.ROW_ID
LEFT OUTER JOIN dbo.S_PROJITEM T13 ON T1.PROJ_ITEM_ID = T13.ROW_ID
LEFT OUTER JOIN dbo.S_SRV_REGN T14 ON T1.SRV_REGN_ID = T14.ROW_ID
LEFT OUTER JOIN dbo.S_SRV_REQ T15 ON T1.SRA_SR_ID = T15.ROW_ID
LEFT OUTER JOIN dbo.S_TIMEZONE T16 ON T14.TIME_ZONE_ID = T16.ROW_ID
LEFT OUTER JOIN dbo.S_OPTY_POSTN T17 ON T1.OPTY_ID = T17.OPTY_ID AND T17.POSITION_ID = '1-D38P'
LEFT OUTER JOIN dbo.S_EVT_ACT_FNX T18 ON T1.ROW_ID = T18.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_FN_APPR T19 ON T18.AMS_ACT_ID = T19.ROW_ID
LEFT OUTER JOIN dbo.S_CASE T20 ON T1.X_CASE_ID = T20.ROW_ID
LEFT OUTER JOIN dbo.S_EMP_PER T21 ON T1.OWNER_PER_ID = T21.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_EVT_ACT_LSX T22 ON T1.ROW_ID = T22.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_EVT_MAIL T23 ON T1.ROW_ID = T23.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_EVT_ACT_SS T24 ON T1.ROW_ID = T24.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_SUSP_ACT T25 ON T1.ROW_ID = T25.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_EVT_MKTG T26 ON T1.ROW_ID = T26.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_EVT_ACT_X T27 ON T1.ROW_ID = T27.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_EVT_CAL T28 ON T1.ROW_ID = T28.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_EVT_CASE T29 ON T1.ROW_ID = T29.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_PARTY T30 ON T1.TARGET_PER_ID = T30.ROW_ID
LEFT OUTER JOIN dbo.S_ACT_EMP T31 ON T1.OWNER_PER_ID = T31.EMP_ID AND T1.ROW_ID = T31.ACTIVITY_ID
LEFT OUTER JOIN dbo.S_PARTY T32 ON T31.EMP_ID = T32.ROW_ID
LEFT OUTER JOIN dbo.S_CONTACT_FNX T33 ON T31.EMP_ID = T33.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_USER T34 ON T31.EMP_ID = T34.PAR_ROW_ID
LEFT OUTER JOIN dbo.S_CON_ADDR T35 ON T1.TARGET_OU_ADDR_ID = T35.ADDR_PER_ID AND T1.TARGET_OU_ID = T35.ACCNT_ID
LEFT OUTER JOIN dbo.S_ADDR_PER T36 ON T1.TARGET_OU_ADDR_ID = T36.ROW_ID
WHERE
((T1.APPT_REPT_REPL_CD IS NULL) AND
((T1.TEMPLATE_FLG != N'Y' AND T1.TEMPLATE_FLG != N'P' OR T1.TEMPLATE_FLG IS NULL) AND (T1.OPTY_ID IS NULL OR T9.SECURE_FLG = N'N' OR T17.OPTY_ID IS NOT NULL)) AND
(T1.PRIV_FLG = 'N' OR T1.PRIV_FLG IS NULL OR T1.OWNER_PER_ID = '1-K6JK')) AND
(T1.ROW_ID = '1-Y95E')
答案 0 :(得分：0)
您在问是否有办法加快转化速度？答案是否定的......否则他们为什么不以此更快地将其作为默认方式？
为什么要转换所有列？你提到它是系统生成的...哪个系统生成它？您可以将表的数据类型更改为正确的类型，然后不需要转换。
答案 1 :(得分：0)
我不相信转换功能可能是原因。它当然需要一些资源，但它肯定需要几分钟。
当转换函数存在时，由于某些原因，你可能会得到一个不同的计划，这可能就是原因。
我建议开始寻找通常会导致查询缓慢的内容：
在您拥有如此庞大数量的表格的情况下，计划创建肯定会超时。您还应该从计划中最左边的操作员处检查。请参阅Grant Fritchey的blog post
如果只比较运行时，可能会得到完全错误的结果，例如因为查询计划已经存在，数据被缓存等等。为了确保只比较时间是正确的，你应该运行checkpoint，{{1 }和dbcc dropcleanbuffers但仅限于测试环境。
答案 2 :(得分：0)
当我使用FORMAT函数将DATETIME转换为字符串时，发现了类似的问题。我尝试对函数进行注释，发现查询显然比以前快。因此，我避免使用FORMAT函数将DATETIME转换为字符串。
就我而言，我想将DATETIME转换为yyyyMM，所以我改用[YEAR(c_date) * 100 + MONTH(c_date)]。
根据我的经验，如果数据不多，可以使用FORMAT函数，但是如果您有数百万条数据记录，则应该避免使用FORMAT函数。