Experimenters regretting throwing away data

Modern experiments, especially in fields like particle physics, often collect far more data than they can process and save. Similarly, I can imagine that historical experimenters were limited by how much they could write down.

Experimenters have to carefully consider which data captured by their sensors to save, and which to throw away, roughly dependent on the theory they are testing.

Are there known examples of instances in which it was later noticed that data which was collected, but thrown away, could have been used? Including instances where a new theory might have been able to be tested by historical data that was thrown away?