R:识别(部分)匹配单元

Question

我有一个有几千行和数百列的表。我需要识别(部分)匹配几个关键字的单元格()，而不是过滤表。到目前为止，我所能找到的只有函数WhichCells {SeuratObject} -唉，我无法让它工作。

例如

Column1 <- c("Temperature", "Water", "Sun", "tree fighter", "rainbow")
Column2 <- c(60.1, 106, 78.6, 21.5, 71)
Column3 <- c("Another cell", NA, "more content", NA, "Thanks!")
    
df <- data.frame(Column1, Column2, Column3)
df

pattern <- c("temp", "content", 55, "heart", "thanks")

希望得到的结果如下所示

pattern   Column    Row
1    temp Column1   1
2 content Column3   3
3  thanks Column3   5

注意:行也可以作为行名返回。

拜托，有人能给我建议怎么解决这个问题吗？非常感谢!

Answer 1

在这里，这是答案的一部分，但您应该做一些调整，以考虑上/下情况：

pattern <- c("Temp", "content", 55, "heart", "Thanks")

detectpattern <- function(df, pattern){
    colindex <- which(apply(df, 1, str_detect, pattern = pattern))
    rowindex <- which(apply(df, 2, str_detect, pattern = pattern))
    return(c(pattern, colindex, rowindex))
}

for (pat in pattern){
    print(detectpattern(df = df, pattern = pat))
}

Answer 2

你可以试试

df %>%
  tibble::rownames_to_column(., var = "Row") %>%
  melt(id.vars = "Row", value.name = "patterns", variable.name = "Column") %>% 
  mutate(check = ((sapply(pattern, function(x) {str_detect(patterns, regex(x, ignore_case = T)) %>% as.numeric} ))%>% rowSums(., na.rm=TRUE))) %>%
  filter(check == 1) %>% select(-check)


  Row  Column     patterns
1   1 Column1  Temperature
2   3 Column3 more content
3   5 Column3      Thanks!