使用一个虚拟示例,我需要完成一个包含隐式缺失值的数据集。这在R中使用tidyr的函数complete是微不足道的。
library(tidyr)
df <- data.frame(Borough = c('Brooklyn', 'Brooklyn', 'Queens'),
Crime = c('Robbery', 'Homicide', 'Drug'),
Count=c(1, 2, 1))
> df
Borough Crime Count
1 Brooklyn Robbery 1
2 Brooklyn Homicide 2
3 Queens Drug 1
#Complete implicit missing values
> complete(df, Borough, Crime, fill=list(Count=0))
Borough Crime Count
1 Brooklyn Drug 0
2 Brooklyn Homicide 2
3 Brooklyn Robbery 1
4 Queens Drug 1
5 Queens Homicide 0
6 Queens Robbery 0但是,在实际数据非常大并且存储在Oracle的SQL表中的情况下,如何使用SQL查询来完成这一任务?
发布于 2017-11-14 16:37:03
交叉连接不同的行政区与犯罪,左加入原来的表,以获得缺失的行与0的罪名。
select b.borough,c.crime,coalesce(t.count,0) as count
from (select distinct borough from tbl) b
cross join (select distinct crime from tbl) c
left join tbl t on t.borough=b.borough and t.crime=c.crimehttps://stackoverflow.com/questions/47290777
复制相似问题