GPUs increasingly accelerate database systems, but query-specific peak performance still often relies on hand-written kernels. Existing LLM kernel benchmarks focus on machine learning operators, leaving irregular, heterogeneous, data-movement-heavy database-style operators untested. We introduce **DataKernelBench**, which translates SQL into validated PyTorch **TorchPlan** programs and evaluates LLMs that optimize either the core tensor-bounded snippet or the full query in CUDA or Triton through execution-guided repair. Across ten proprietary and open-weight models on **TPC-H SF10** with an H100 GPU, the strongest full-query CUDA configuration achieves **2.11× speedup over ‘torch.compile‘** at full pass rate. We find that higher-performing implementations commonly use kernel fusion and execution-strategy changes, stronger models benefit most from full-query specialization, and workload context matters more than hardware context. To handle data larger than GPU memory, we extend TorchPlan with Dask-cuDF for on-demand partition loading on **TPC-H SF100** with four H100 GPUs, achieving **2.54×** speedup.

**Project page:** https://kerneldf.github.io/datakernelbench