This can be easily accomplished by using the filter() verb from dplyr. You can read more about it here: https://dplyr.tidyverse.org/reference/filter.html
For pattern matching, you can use regular expressions (regex for short). Learn more about them here: https://regexone.com/
library(dplyr, warn.conflicts = FALSE)
#> Warning: package 'dplyr' was built under R version 3.6.3
library(stringr, warn.conflicts = FALSE)
df <- tribble(~ row, ~ col, ~ value,
1, 0, "ID",
1, 1, 12,
1, 2, 12,
1, 3, 4,
1, 4, NA)
filter(df, str_detect(value, "[0-9]+"))
#> # A tibble: 3 x 3
#> row col value
#> <dbl> <dbl> <chr>
#> 1 1 1 12
#> 2 1 2 12
#> 3 1 3 4
Created on 2020-03-30 by the reprex package (v0.3.0)