I have two dataframes (df1 and df2) that look like this:
df1:
df2:
My aim is to change the IDs of df2 to geo_accession using df1 as reference:
Example: SYMBOL GSM4187205 ......
But I have encountered two problems:
df2.mod <- df2 %>%
gather(key = "samples", value = "counts", -SYMBOL) %>%
mutate(samples = gsub("X", "", samples)) %>%
inner_join(., df1, by = c('samples' = 'title')) %>%
spread(key = 'geo_accession', value = 'counts') %>%
column_to_rownames(var = 'SYMBOL')
But it doesn't work because the number of samples and title variables don't match.
Error: <0 rows> (o 0- extensión row.names)
Thank you for your help,
gsub might be removing more than you need, see example:
id <- c("X123", "X234_A", "X234_X")
gsub("X", "", id)
# [1] "123" "234_A" "234_"
gsub("^X", "", id)
# [1] "123" "234_A" "234_X"
Either fix the gsub line or when you are reading in the data set check.names = FALSE:
df2 <- read.table("myfile.txt", header = TRUE, check.names = FALSE)
Hi,
Adding when I am reading the dataset: check.names = FALSE, it's works! The names come out without the X but, the final result after the "inner_join" is:
[1] SYMBOL samples counts geo_accession
<0 rows> (o 0- extensión row.names)
I think that it's the number of the variables for name ID: in df2 is 552 and in df1 is 587. How do I know and remove excess variables from df1?
Log in to answer this question.
Can you confirm that values in header of dataframe 2 correspond to column 2 of dataframe 1?
Hi,
I confirm. I I manually checked the first 10 with the full dataset. Both files come from the same data: one is the metadata dataset and the other is the gene expression dataset.