Dremel. Interactive Analysis of Web-Scale Datasets

• Tags:: 📜Papers
• File:
  • {{pdf: https://firebasestorage.googleapis.com/v0/b/firescript-577a2.appspot.com/o/imgs%2Fapp%2Fdr___mario_second_brain%2FaisXSatFRu.pdf?alt=media&token=b7d759d5-33a1-4a59-af3d-d8ee2b8890ba}}
• Highlights:
  • Big Query is based on this.
  • So it seems that most of the work went to make repeated and nesting records efficient because, at scale, “normalizing and recombining” became prohibitive, and “the data used in web and scientific computing is often non-relational”.
  • But the thing is that, storing data like that makes retrieval costly, so it only makes sense if you cannot work with normalized data.