Press "Enter" to skip to content

Loading Data from CSVs with Inconsistent Quoted Identifiers

Dave Mason has some fun with loading data from files:

BCP and OPENROWSET are long-lived SQL Server options for working with data in external files. I’ve blogged about OPENROWSET, including a recent article showing a way to deal with quoted data. One of the shortcomings I’ve never been able to overcome is an inconsistent data file with data fields in some rows enclosed in double quotes, but not all. I’ve never found a way around this limitation.

Let’s demonstrate with BCP. Below is a sample data file I’ll attempt to load into a SQL Server table. Note the data fields highlighted in yellow, which are enclosed in double quotes and contain the field terminator , (comma) character. For reference, the file is also available on Github.

I get unduly frustrated with the implementations of various data loaders around SQL Server and how they handle quoted identifiers differently. And don’t get me started on PolyBase.