Export Data CSV fails when field contains a # character

While exporting some data via the Export Data CSV feature, I've noticed incomplete files being generated. After doing some more testing, I've found out the writing of the file would end prematurely exactly at the point where a `#` character would be found in the field that was being written, e.g. a table containing:
```
column_a,column_b,column_c
1,blah blah,blah
2,blah #blah,blah
3,blah,blah blah
```
will produce a file like:
```
column_a,column_b
1,blah blah blah
2,blah
```
Please note that the next column and any other possible record would be skipped entirely.

Start at the Export Data CSV entry point and reproduce the failure with a field containing a # character. Verify that the generated CSV preserves the remaining columns and records after the character, including the example rows shown in the issue.

Written by the indexing model from the issue text.