bugfix in ogs5_read_tecplot_domain()
I think this part
#=== get point data
suppressWarnings(
suppressMessages(
df <- readr::read_delim(filename,
" ", escape_double = FALSE, col_names = FALSE,
trim_ws = TRUE, skip = 3)))
should be changed to something like:
#=== get point data
df <- readr::read_delim(fileurl,
" ", escape_double = T, col_names = FALSE,
trim_ws = TRUE, skip = 3, col_types = cols(.default = "c"))
It is not a problem for the output of ex1 somehow, but for some other benchmarks (e.g. H/Theis/LF_Theis_3D/LF_Theis_3D_domain_LIQUID_FLOW_hex.tec), if it's not specified that the columns will be read in as characters, readr will read them in as double:
Then in the following lines,
# extract all lines with elements in first column that do not match pattern
df<-df %>%
dplyr::filter(
!is.na(stringr::str_extract(X1,"\\d+\\.\\d+e\\+\\d+"))|!is.na(stringr::str_extract(X1,"\\d+\\.\\d+e\\-\\d+"))
)
stringr will not be able to extract the lines