See ?write.table documentation for "qmethod", which needs to be set
to "double" to write the data correctly for csv. This is most easily
done by using read.csv and write.csv, for which this is the default.
Ergo, the following :
(using your example)
write.csv(db, file = "test", row.names = FALSE)
## Note: Set row.names = FALSE so an extra column of numeric row names
won't be added.
db2 <- read.csv("test") ## read it back in
identical(db2, db)
[1] TRUE
Apologies if I have misunderstood and this does not solve your problem.
Cheers,
Bert
On Sun, Sep 14, 2025 at 10:46?AM Patrick Giraudoux
<patrick.giraudoux at univ-fcomte.fr> wrote:
Dear listers, previous plain text was still more messy,
Here a trial hopefully better...
I encountered an issue with a CSV file that was imported correctly but
could not be re-imported correctly after being written with R. This is
probably because geographical coordinates were imported as character in
degrees, minutes and seconds (DMS), which includes " (quotation mark)
for the seconds.
Below a reproducible example:
db <- structure(list(lon = c(6.228561, 6.22532, 6.2260499999999999,
6.2267789999999996, 6.2224659999999998, 6.2209430000000001), latdms =
c("47?12'28.36\"N", "47?12'33.46\"N", "47?12'28.37\"N",
"47?12'27.48\"N", "47?12'31.31\"N", "47?12'33.15\"N"), londms =
c("6?13'42.82\"E", "6?13'31.15\"E", "6?13'33.78\"E", "6?13'36.40\"E",
"6?13'20.88\"E", "6?13'15.39\"E"), fusutmn = c(32L, 32L, 32L, 32L, 32L,
32L)), row.names = c(NA, 6L), class = "data.frame")
write.table(db, file = "db.txt", row.names = FALSE, quote = FALSE, sep =
"\t")
db_import<-read.delim("db.txt")
As you can see it, latdms and londms are now collapsed and all the
columns on the right have shifted to the left. I get the same issue
with ; as a separator.
I could not find a workaround...
Any hint appreciated,
Patrick