Question

我想合并两个数据框：

df1：

Date         Company    Return
1988-09-30   BELSHIPS   0.087
1988-10-31   BELSHIPS   0.021
1988-11-30   BELSHIPS   0.015
1988-12-30   BELSHIPS   -0.048
1988-09-30   GOODTECH   0.114
1988-10-31   GOODTECH   0.074
1988-11-30   GOODTECH   NA
1988-12-30   GOODTECH   NA
1988-09-30   LABOREMUS  -0.014
1988-10-31   LABOREMUS  0.024
1988-11-30   LABOREMUS  0.017
1988-12-30   LABOREMUS  0.021

df2：

Company
BELSHIPS
BIK BOK
FARSTAD SHIPPING
GOODTECH
GYLDENDAL

我想按公司合并这两个数据框，但我只想包括那些没有NA的公司。因此，新数据框应如下所示：

df3：

Date         Company    Return
1988-09-30   BELSHIPS   0.087
1988-10-31   BELSHIPS   0.021
1988-11-30   BELSHIPS   0.015
1988-12-30   BELSHIPS   -0.048

只包括公司BELSHIPS，因为GOODTECH在回报中有NA，而LABOREMUS不包含在df2中。

我尝试了df3 <- merge(df2, df1[!is.na(df1$Return)], by = "Company")这不起作用，因为它只省略了具有NA的行，而不是整个公司。

有关如何解决这个问题的任何建议吗？

Answer 1

Base R解决方案：

# Select companies that have NA
# You can also use unique on this
foo <- df1$Company[is.na(df1$Return)]
# Subset data frame where Company is within df2 and doesn't have NA
subset(df1, Company %in% df2$Company & !Company %in% foo)

#         Date  Company Return
# 1 1988-09-30 BELSHIPS  0.087
# 2 1988-10-31 BELSHIPS  0.021
# 3 1988-11-30 BELSHIPS  0.015
# 4 1988-12-30 BELSHIPS -0.048

Answer 2

测试数据：

df2 = data.frame(Company = c('BELSHIPS','GOODTECH'))                                                                                                                                                                             
df1 = data.frame(Company = c('BELSHIPS','BELSHIPS','BELSHIPS','GOODTECH','GOODTECH','GOODTECH','LABOREMUS','LABOREMUS','LABOREMUS'),Return = c(1,2,3,1,NA,NA,3,4,5) )

使用which()和unique()抓取NA行的公司：

df3<-merge(df2, df1[df1$Company!=unique(df1[which(is.na(df1$Return)),'Company']),], by = 'Company')

Answer 3

您也可以使用dplyr：

执行此操作

df2 %>%
  left_join(df1, by = "Company") %>%
  group_by(Company) %>% 
  filter(sum(is.na(Return)) == 0)

给你：

# A tibble: 4 x 3
# Groups:   Company [1]
   Company       Date Return
     <chr>     <fctr>  <dbl>
1 BELSHIPS 1988-09-30  0.087
2 BELSHIPS 1988-10-31  0.021
3 BELSHIPS 1988-11-30  0.015
4 BELSHIPS 1988-12-30 -0.048

Answer 4

只需将其合并，然后使用函数na.omit(merged df)

合并两个数据帧，但仅包括没有NA的变量

4 个答案: