R 等效于 SQL 中的情况？

Question

I have been using T-SQL for a while but I now have to make transition to R...

select case when date_column_1 < '20160801' then date_column_1 else '20160801' end as date_column_1, case when date_column_2 < '99991231' then '20190701' else date_column_2 end as date_column_2 from table

Also, apologies in advance with my crappy stackoverflow formatting skills.

Answer 1

Since you have only one condition to check for each column, we can use ifelse here. If there are multiple conditions you could check case_when from dplyr .

transform(table, date_column_1 = ifelse(date_column_1 < as.Date('2016-08-01'), 
                                        date_column_1, as.Date('2016-08-01')), 
                 date_column_2 = ifelse(date_column_2 < as.Date('9999-12-31'),
                                        as.Date('2019-07-01'), date_column_2))

Assuming date_column_1 and date_column_2 are of Date class, you cannot directly compare dates with literal string ( "20160801" ) in R, so we convert the dates to compare to Date class as well. Moreover, for date_column_2 do you mean to replace all values with the date "2019-07-01" ?

---

As mentioned by @G. Grothendieck ifelse would convert the dates to numeric. We can use if_else from dplyr which is type-strict and will maintain the dates.

library(dplyr)

table %>%
   mutate(date_column_1 = if_else(date_column_1 < as.Date('2016-08-01'), 
                            date_column_1, as.Date('2016-08-01')), 
          date_column_2 = if_else(date_column_2 < as.Date('9999-12-31'),
                            as.Date('2019-07-01'), date_column_2)))