如何从R中的数据框中删除“很多”行

Question

I tried all the similar posts but none of the answers seemed to work for me. I want to delete 8500+ rows (by rowname only) from a dataframe with 27,000+. The other columns are completely different, but the smaller dataset was derived from the larger one, and just looking for names shows me that whatever I look for from smaller df it is present in larger df. I could of course do this manually (busy work for sure!), but seems like there should be a simple computational answer.

I have tried:

fordel<-df2[1,]

df3<-df1[!rownames(df1) %in% fordel

l1<- as.vector(df2[1,])

df3<- df1[1-c(l1),]

and lots of other crazy ideas! Here is a smallish example: df1:

Ent_gene_id clone57_RNA clone43_RNA_2   clone67_RNA clone55_RNA
ENSMUSG00000000001.4    10634   6954    6835    6510
ENSMUSG00000000003.15   0       0       0       0
ENSMUSG00000000028.14   559     1570    807     1171
ENSMUSG00000000031.15   5748    174     4103    146
ENSMUSG00000000037.16   37      194     49      96
ENSMUSG00000000049.11   0       3       1       0
ENSMUSG00000000056.7    1157    1125    806     947
ENSMUSG00000000058.6    75      304     123     169
ENSMUSG00000000078.6    4012    4391    5637    3854
ENSMUSG00000000085.16   381     560     482     368
ENSMUSG00000000088.6    2667    4777    3483    3450
ENSMUSG00000000093.6    3       48      41      22
ENSMUSG00000000094.12   23      201     102     192

df2

structure(list(base_mean = c(7962.408875, 947.1240794, 43.76698418 ), log2foldchange = c(-0.363434063, -0.137403759, -0.236463207 ), lfcSE = c(0.096816743, 0.059823215, 0.404929452), stat = c(-3.753834854, -2.296830066, -0.583961493)), row.names = c("ENSMUSG00000000001.4", "ENSMUSG00000000056.7", "ENSMUSG00000000093.6"), class = "data.frame")

I want to delete from df1 the rows corresponding to the rownames in df2. Tried to format it, but seems no longer formatted... oh well....

Suggestions really appreciated!

Answer 1

You mentioned row names but your data does not include that, so I'll assume that they really don't matter (or exist). Also, your df2 has more column headers than columns, not sure what's going on there ... so I'll ignore it.

Data

df1 <- structure(list(Ent_gene_id = c("ENSMUSG00000000001.4", "ENSMUSG00000000003.15", 
"ENSMUSG00000000028.14", "ENSMUSG00000000031.15", "ENSMUSG00000000037.16", 
"ENSMUSG00000000049.11", "ENSMUSG00000000056.7", "ENSMUSG00000000058.6", 
"ENSMUSG00000000078.6", "ENSMUSG00000000085.16", "ENSMUSG00000000088.6", 
"ENSMUSG00000000093.6", "ENSMUSG00000000094.12"), clone57_RNA = c(10634L, 
0L, 559L, 5748L, 37L, 0L, 1157L, 75L, 4012L, 381L, 2667L, 3L, 
23L), clone43_RNA_2 = c(6954L, 0L, 1570L, 174L, 194L, 3L, 1125L, 
304L, 4391L, 560L, 4777L, 48L, 201L), clone67_RNA = c(6835L, 
0L, 807L, 4103L, 49L, 1L, 806L, 123L, 5637L, 482L, 3483L, 41L, 
102L), clone55_RNA = c(6510L, 0L, 1171L, 146L, 96L, 0L, 947L, 
169L, 3854L, 368L, 3450L, 22L, 192L)), class = "data.frame", row.names = c(NA, 
-13L))
df2 <- structure(list(Ent_gene_id = c("ENSMUSG00000000001.4", "ENSMUSG00000000056.7", 
"ENSMUSG00000000093.6"), base_mean = c(7962.408875, 947.1240794, 
43.76698418), log2foldchange = c(-0.36343406, -0.137403759, -0.236463207
), pvalue = c(0.00017415, 0.021628466, 0.55924622)), class = "data.frame", row.names = c(NA, 
-3L))

Base

df1[!df1$Ent_gene_id %in% df2$Ent_gene_id,]
#              Ent_gene_id clone57_RNA clone43_RNA_2 clone67_RNA clone55_RNA
# 2  ENSMUSG00000000003.15           0             0           0           0
# 3  ENSMUSG00000000028.14         559          1570         807        1171
# 4  ENSMUSG00000000031.15        5748           174        4103         146
# 5  ENSMUSG00000000037.16          37           194          49          96
# 6  ENSMUSG00000000049.11           0             3           1           0
# 8   ENSMUSG00000000058.6          75           304         123         169
# 9   ENSMUSG00000000078.6        4012          4391        5637        3854
# 10 ENSMUSG00000000085.16         381           560         482         368
# 11  ENSMUSG00000000088.6        2667          4777        3483        3450
# 13 ENSMUSG00000000094.12          23           201         102         192

dplyr

dplyr::anti_join(df1, df2, by = "Ent_gene_id")
#              Ent_gene_id clone57_RNA clone43_RNA_2 clone67_RNA clone55_RNA
# 1  ENSMUSG00000000003.15           0             0           0           0
# 2  ENSMUSG00000000028.14         559          1570         807        1171
# 3  ENSMUSG00000000031.15        5748           174        4103         146
# 4  ENSMUSG00000000037.16          37           194          49          96
# 5  ENSMUSG00000000049.11           0             3           1           0
# 6   ENSMUSG00000000058.6          75           304         123         169
# 7   ENSMUSG00000000078.6        4012          4391        5637        3854
# 8  ENSMUSG00000000085.16         381           560         482         368
# 9   ENSMUSG00000000088.6        2667          4777        3483        3450
# 10 ENSMUSG00000000094.12          23           201         102         192

---

Edit : same thing but with row names:

# update my df1 to change Ent_gene_id from a column to rownames
rownames(df1) <- df1$Ent_gene_id
df1$Ent_gene_id <- NULL
# use your updated df2 (from dput)
# df2 <- structure(...)
df1[ !rownames(df1) %in% rownames(df2), ]
#                       clone57_RNA clone43_RNA_2 clone67_RNA clone55_RNA
# ENSMUSG00000000003.15           0             0           0           0
# ENSMUSG00000000028.14         559          1570         807        1171
# ENSMUSG00000000031.15        5748           174        4103         146
# ENSMUSG00000000037.16          37           194          49          96
# ENSMUSG00000000049.11           0             3           1           0
# ENSMUSG00000000058.6           75           304         123         169
# ENSMUSG00000000078.6         4012          4391        5637        3854
# ENSMUSG00000000085.16         381           560         482         368
# ENSMUSG00000000088.6         2667          4777        3483        3450
# ENSMUSG00000000094.12          23           201         102         192