Builds a fresh harbour_dtable from CSV files, one table per file.
This is not the inverse of hb_write_csv(): it produces a new base
with new identifiers, and only text, number and checkbox columns.
Nothing that a CSV cannot carry is recovered.
Value
A harbour_dtable, with the components hb_read_dtable()
describes, holding one table per file.
Details
Columns are typed the way utils::read.csv() types them, and quoting a
value does not change that:
A column of numbers becomes
number, and one holding onlyT,F,TRUEorFALSEbecomescheckbox.A value written
NAis read as missing.Dates and date-times arrive as text.
Some exceptions avoid losing data. A column holding an integer with a
leading zero, such as "0001", a value with a leading plus, such as
the telephone number "+4917612345678", or an integer of more than 15
digits stays text, rather than losing the zeros, the plus or digits,
and so does one holding a code such as "12E", which
utils::read.csv() reads as 12 in R before 4.4.1. A column with no
values at all is text, not a checkbox column that would read FALSE
in every row.
The files must be UTF-8, with or without a byte-order mark. A file in
another encoding is refused, naming the first line that is not valid
UTF-8. In a multibyte session that is not UTF-8, such as Shift-JIS or
GBK, the text goes through that session's encoding, so characters it
lacks are lost; read such files in a UTF-8 session. A column named
_id, such as the one hb_write_csv() writes,
is dropped: SeaTable keeps that name for the row identifier, and the new
base has identifiers of its own. One that holds anything but SeaTable
row ids, such as a database's keys, is dropped with a warning.
Examples
csv <- tempfile(fileext = ".csv")
utils::write.csv(data.frame(x = 1:2, y = c("a", "b")), csv,
row.names = FALSE)
hb_read_csv(c(Measurements = csv))
#>
#> ── <harbour_dtable> ────────────────────────────────────────────────────────────
#> • base : "harbouR base"
#> • tables : 1
#> • rows : 2
#> • assets : 0
#>
#> - Measurements (2 cols, 2 rows)
unlink(csv)