Question

我需要提高此代码的效率（请参阅下文）。代码正在做的是更新主数据集＆＃39;合并＆＃39;来自＆＃39;的价值观＆＃39;只有在观察到的情况下才能巩固＆＃39;不见了。这将通过并更新16个变量。

有没有办法提高效率，例如在16个变量中创建一个循环？

proc sql;
update index.consolidated as a 
set GRANTS_3YP = case when a.GRANTS_3YP ^= . then a.GRANTS_3YP else (select GRANTS_3YP from index.means as b where a.LGA = b.LGA) end,
    GRANTS_3Y = case when a.GRANTS_3Y ^= . then a.GRANTS_3Y else (select GRANTS_3Y from index.means as b where a.LGA = b.LGA) end,
    RI_GEN = case when a.RI_GEN ^= . then a.RI_GEN else (select RI_GEN from index.means as b where a.LGA = b.LGA) end,
    FIFO = case when a.FIFO ^= . then a.FIFO else (select FIFO from index.means as b where a.LGA = b.LGA) end,
    YU_3Y = case when a.YU_3Y ^= . then a.YU_3Y else (select YU_3Y from index.means as b where a.LGA = b.LGA) end,
    POP_FLOWS = case when a.POP_FLOWS ^= . then a.POP_FLOWS else (select POP_FLOWS from index.means as b where a.LGA = b.LGA) end,
    COL = case when a.COL ^= . then a.COL else (select COL from index.means as b where a.LGA = b.LGA) end,
    IC = case when a.IC ^= . then a.IC else (select IC from index.means as b where a.LGA = b.LGA) end,
    DISTANCE = case when a.DISTANCE ^= . then a.DISTANCE else (select DISTANCE from index.means as b where a.LGA = b.LGA) end,
    MR_3Y = case when a.MR_3Y ^= . then a.MR_3Y else (select MR_3Y from index.means as b where a.LGA = b.LGA) end,
    MP = case when a.MP ^= . then a.MP else (select MP from index.means as b where a.LGA = b.LGA) end,
    DP = case when a.DP ^= . then a.DP else (select DP from index.means as b where a.LGA = b.LGA) end,
    RENT = case when a.RENT ^= . then a.RENT else (select RENT from index.means as b where a.LGA = b.LGA) end,
    DEATH = case when a.DEATH ^= . then a.DEATH else (select DEATH from index.means as b where a.LGA = b.LGA) end,
    MENTAL = case when a.MENTAL ^= . then a.MENTAL else (select MENTAL from index.means as b where a.LGA = b.LGA) end,
    AGE = case when a.AGE ^= . then a.AGE else (select AGE from index.means as b where a.LGA = b.LGA) end;
quit;

Answer 1

您的查询将花费太多时间，因为您正在同时更新16条记录，然后您在每个SET中使用子查询。

我的建议是使用JOIN（INNER / LEFT / RIGHT），似乎每个SET中的子查询包含相同的表index.means。只需加入一个标识即可加入您的表格。

PROC SQL UPDATE - 效率

1 个答案: