datatable.iread()¶
This function is similar to fread()
, but allows reading
multiple sources at once. For example, this can be used when the
input is a list of files, or a glob pattern, or a multi-file archive,
or multi-sheet XLSX file, etc.
Parameters¶
...
Most parameters are the same as in fread()
. All parse
parameters will be applied to all input files.
"warn"
| "raise"
| "ignore"
| "store"
What action to take when one of the input sources produces an
error. Possible actions are: "warn"
– each error is converted
into a warning and emitted to user, the source that produced the
error is then skipped; "raise"
– the errors are raised
immediately and the iteration stops; "ignore"
– the erroneous
sources are silently ignored; "store"
– when an error is
raised, it is captured and returned to the user, then the iterator
continues reading the subsequent sources.
Iterator[Frame]
| Iterator[Frame|Exception]
The returned object is an iterator that produces Frame
s.
The iterator is lazy: each frame is read only as needed, after the
previous frame was “consumed” by the user. Thus, the user can
interrupt the iterator without having to read all the frames.
Each Frame
produced by the iterator has a .source
attribute that describes the source of each frame as best as
possible. Each source depends on the type of the input: either a
file name, or a URL, or the name of the file in an archive, etc.
If the errors
parameter is "store"
then the iterator may
produce either Frames or exception objects.