Match MySQL `float64` to `float32` conversion precision loss

MySQL drops more precision for float32 conversions.
Dolt and Postgres:
```
tmp/main*> select cast(20010203123456 as float);
+-------------------------------+
| cast(20010203123456 as float) |
+-------------------------------+
| 2.0010202e+13 |
+-------------------------------+
1 row in set (0.00 sec)
```

MySQL:
```
mysql> select cast(20010203123456 as float);
+-------------------------------+
| cast(20010203123456 as float) |
+-------------------------------+
| 20010200000000 |
+-------------------------------+
1 row in set (0.00 sec)
```

Start by reproducing the SQL CAST example for 20010203123456 and locate the float64-to-float32 conversion path responsible for the result. Add a regression test for the conversion, and consider the work complete when the result matches the expected MySQL precision and the test passes.

Written by the indexing model from the issue text.