将STATA语法转换为R语法

Question

我在将一些STATA代码翻译成R代码时遇到了一些麻烦:

占据代码:

gen joint_gpw = sbud_jpw * q44 if sbud_jpw < 888 & q44 < 888 
gen sbud_gpw_all = sbud_gpw if sbud_gpw < 888 
replace sbud_gpw_all = q31 if sbud_gpw_all ==. & q31 < 888 
replace sbud_gpw_all = joint_gpw if sbud_gpw_all ==. & joint_gpw !=. 
replace sbud_gpw_all = 888 if q16_1 == 0 & sbud_gpw_all ==. 
replace sbud_gpw_all = 888 if (sbud_gpw == 888 & q31 == 888 & sbud_jpw == 888 & q44 == 888) & sbud_gpw_all ==.  
replace sbud_gpw_all = 999 if (sbud_gpw == 999 | q31 == 999 | sbud_jpw == 999 | q44 == 999  | (q44 !=. & sbud_jpw == 888)) & sbud_gpw_all ==. 

下面是我试过的R代码:

dat%>%
dplyr::mutate(joint_gpw = ifelse((sbud_jpw<888 & q44<888),sbud_jpw * q44,NA))%>%
dplyr::mutate(sbud_gpw_all = ifelse(sbud_gpw < 888,sbud_gpw,NA))%>%
dplyr::mutate(sbud_gpw_all = ifelse((sbud_gpw_all= NA & q31<888),q31,NA))%>%
dplyr::mutate(sbud_gpw_all = ifelse((sbud_gpw_all = NA & joint_gpw != NA),joint_gpw,NA))%>%
dplyr::mutate(sbud_gpw_all) = ifelse((q16_1 = 0 & sbud_gpw_all = NA),888,NA)%>%
dplyr::mutate(sbud_gpw_all) = ifelse((sbud_gpw = 888 & q31 = 888 & sbud_jpw = 888 & q44 = 888) & sbud_gpw_all = NA,888,NA)%>%
dplyr::mutate(sbud_gpw_all) = ifelse(((sbud_gpw = 999 | q31 = 999 | sbud_jpw = 999 | q44 = 999  | (q44 != NA & sbud_jpw == 888)) & sbud_gpw_all = NA)),999,NA)

之前出现的错误:

Error: unexpected '=' in:
"  dplyr::mutate(sbud_gpw_all) = ifelse((q16_1 = 0 & sbud_gpw_all = NA),888,NA)%>%
dplyr::mutate(sbud_gpw_all) = ifelse((sbud_gpw = 888 & q31 = 888 & sbud_jpw = 888 & q44 = 888) & sbud_gpw_all ="

我想知道这两组码是否等价?我非常感谢所有的帮助!谢谢! !

Answer 1

错误源于最后三行sbud_gpw_all后面的右括号)

另外，尽管没有抛出错误，但您正在用每个突变覆盖sbud_gpw_all。我不知道Stata，你没有提供一个最小的可重复的例子，但我有一种感觉，你的代码可以这样工作:

dat %>%
mutate(
joint_gpw = if_else(sbud_jpw < 888 & q44 < 888, sbud_jpw * q44, NA_real_),
sbud_gpw_all = case_when(
sbud_gpw < 888 ~ sbud_gpw,
q31 < 888 ~ q31,
!is.na(joint_gpw) ~ joint_gpw,
q16_1 == 0 ~ 888,
sbud_gpw == 888 & q31 == 888 & sbud_jpw == 888 & q44 == 888 ~ 888,
sbud_gpw == 999 | q31 == 999 | sbud_jpw == 999 | q44 == 999 | (!is.na(q44) & sbud_jpw == 888) ~ 999
)
)

这将首先使用sbud_jpw < 888 & q44 < 888创建dplyr::if_else()列joint_gpw。之后，有一组条件(在~之前)被依次检查。第一个匹配行，提供值(在~操作符后面)。

请注意，正如Sotos在评论中指出的那样，R中的NAs是用is.na(x)检查的，而不是用==/!=，因为它们总是返回NA。我省略了大多数行的NA检查，因为这些检查隐含在case_when()的顺序性质中——只要一个条件匹配，后面的条件就不再求值。NA_real_为数字NA值。使用if_else()和case_when()，您必须明确数据类型。