feat: support scientific notation in parse_float_as_decimal
#8494
Conversation
| @@ -206,6 +206,18 @@ select 123456789.0123456789012345678901234567890 | |||
| statement error SQL error: ParserError\("Cannot parse 1234567890123456789012345678901234567890 as i128 when building decimal: number too large to fit in target type"\) | |||
| select -123456789.0123456789012345678901234567890 | |||
|
|
|||
| # scientific notation | |||
| query RRRR | |||
| select 1.234e2, 1.234e-2, -1.234e2, -1.234e-2 | |||
There was a problem hiding this comment.
1.234e-2 is not -123.4 I don't think 🤔
DataFusion CLI v33.0.0
❯ select 1.234e-2;
+------------------+
| Float64(0.01234) |
+------------------+
| 0.01234 |
+------------------+
datafusion/sql/src/expr/value.rs
Outdated
| ))) | ||
| })?; | ||
|
|
||
| let replaced_str = big_decimal.to_string().replace('.', ""); |
There was a problem hiding this comment.
Once you have parsed the data into a BigDecimal, I think you could just convert to a i128 directly which would be more performant and I think potentially more correct
Perhaps you could use
https://docs.rs/bigdecimal/latest/bigdecimal/struct.BigDecimal.html#method.fractional_digit_count to get the scale
And then compute the precision and the i128 from
https://docs.rs/bigdecimal/latest/bigdecimal/struct.BigDecimal.html#method.into_bigint_and_exponent
|
I don't think this PR is waiting on feedback, so marking it as a draft. Please mark it as ready for review when it is ready for another look 🙏 |
|
Thank you for your contribution. Unfortunately, this pull request is stale because it has been open 60 days with no activity. Please remove the stale label or comment or this will be closed in 7 days. |
Which issue does this PR close?
Closes #8486
Rationale for this change
What changes are included in this PR?
Are these changes tested?
Are there any user-facing changes?