scruz-denodo opened a new pull request, #1407: URL: https://github.com/apache/arrow-java/pull/1407
## What's Changed `ArrowFlightJdbcFloat4VectorAccessor` no longer exposes float-widening artifacts when a `Float4Vector ` is read as a `double` or a `decimal`. - `getDouble()` (and therefore `getObject(Double.class)`) now returns `Double.parseDouble(Float.toString(value))` instead of `(double) value`. A `0.1f` cell now gives `0.1,` not` 0.10000000149011612`. - `getBigDecimal() `and `getBigDecimal(int scale)` now use `new BigDecimal(Float.toString(value))` instead of `BigDecimal.valueOf(value)`, which widened the float to a double first. - `NaN` and `±Infinity` are still returned by `getDouble()`. `getBigDecimal()` still throws `SQLException` for them, and `null `cells still return `0.0` / `null`. **Tests (ArrowFlightJdbcFloat4VectorAccessorTest)**: - `getDouble `and `getBigDecimal `expectations now use Float.toString. - A new test checks that `(float) getDouble() == getFloat()`. - A new test covers 0.1f, -0.1f, 0.2f, 0.3f, 1.1f, 123.456f, 3.4e-5f and -0.0f through `getDouble`, `getObject(Double.class)`, `getBigDecimal()` and `getBigDecimal(9)`. - A 0.1f test checks 0.1, 0.1 and 0.100000000. - A `NaN getDouble` test is added. Closes #1406. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
