Skip to content

Commit 289f907

Browse files
committed
Update formats.md
1 parent cf2f3b1 commit 289f907

File tree

1 file changed

+1
-1
lines changed

1 file changed

+1
-1
lines changed

docs/src/man/formats.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -204,7 +204,7 @@ julia> setformat!(ds, 1:3 => fmt1)
204204

205205
## `format` validation
206206

207-
InMemoryDatasets doesn't validate the supplied `format` until it needs to use the formatted values for an operation, in that case, if the supplied `format` is not a valid `format`, InMemoryDatasets will throw errors. Also it is important to note that InMemoryDatasets is not aware of changing the definition of a `format` by users, thus, changing the definition of a function which is used as a `format` during a workflow may have some side effects. For example if a data set is `groupby!` with `mapformats = true` option, changing the definition of the formats destroys the sorting order of the data set, but InMemoryDatasets is unaware of this, so, it is the user responsibility to remove the invalid formats in these situations.
207+
InMemoryDatasets doesn't validate the supplied `format` until it needs to use the formatted values for an operation, in that case, if the supplied `format` is not a valid `format`, InMemoryDatasets will throw errors. Also it is important to note that InMemoryDatasets is not aware of changing the definition of a `format` by users, thus, changing the definition of a function which is used as a `format` during a workflow may have some side effects. For example if a data set is `groupby!` with `mapformats = true` option, changing the definition of the formats invalidates the sorting order of the data set, but InMemoryDatasets is unaware of this, so, it is the user responsibility to remove the invalid formats in these situations.
208208

209209
In the following examples we demonstrate some scenarios which end up with an invalid `format`, and provide some remedies to fix the issues. Nevertheless, note that supplying an invalid `format` will not damage a data set and a simple call to `removeformat!` can be helpful to recover the original data set.
210210

0 commit comments

Comments
 (0)