Github user MaxGekk commented on a diff in the pull request: https://github.com/apache/spark/pull/22237#discussion_r234792965 --- Diff: R/pkg/tests/fulltests/test_sparkSQL.R --- @@ -1694,7 +1694,7 @@ test_that("column functions", { df <- as.DataFrame(list(list("col" = "{\"date\":\"21/10/2014\"}"))) schema2 <- structType(structField("date", "date")) s <- collect(select(df, from_json(df$col, schema2))) - expect_equal(s[[1]][[1]], NA) + expect_equal(s[[1]][[1]]$date, NA) --- End diff -- Do you mean this particular line or in general? This line was changed because in the `PERMISSIVE` mode we usually return a `Row` with null fields that we wasn't able to parse instead of just `null` for whole row. In general, to fully support the `PERMISSIVE` mode without any excuses when uniVocity parser cannot detect any JSON tokens on root level. We switched to `FailureSafeParser` in `from_json` and `PERMISSIVE` as the default mode, recently there #22237. Previously `from_json` didn't support any modes comparing to JSON datasource.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org