1. 8

I am doing research on E. F. Codd’s work. I wonder since then, 1970, what is the most tables you have used in a real database? Thank you in advance.

  1.  

  2. 4

    I’m not sure the size of the largest database I’ve created myself, but the largest I’ve used would be an SAP ECC system. Having a quick look, one of the systems I have access to has 28,247 tables (30,857 indexes). This is a slightly older system (ECC 6.0 EhP4) - a more up-to-date system would no doubt have several thousand more tables. Total size of the system in question is roughly 380GB.

    1. 3

      Dozens of thousands, though it’s postgresql, which requires separate tables for every partition of a partitioned table. Barring partitions, it’s probably only ten thousand or so. The largest database with that schema that I manage is in the 25TB range.

      1. 2

        Only worked at one company, primarily on one app.

        There are 156 tables in that app.

        Edit: Sorry, I’m new here, didn’t know the formatting would change “156.” to “1.”

        1. [Comment removed by author]

          1. 2

            Our wordpress platform (wordpress MU) has north of 46000 tables. It seem that WordPress thought it a sound design decision to have distinct tables for every single blog…

            1. 1

              107

              1. 1

                46

                1. 1

                  From the few DBs I have on this machine, 148 is the most tables in any one of them.

                  1. 1

                    It wasn’t my database, but I saw an application once, for a satellite tv provider, that dynamically created a db table (not row) for every single customer. So there were hundreds of thousands of tables in the app.

                    1. 1

                      Can’t say I’ve ever had any projects with hundreds or thousands of tables with different schemata, but one distributed metrics storage system I worked on had a big database sharded across probably 500+ tables worldwide. The first layer of sharding was regional and depended on where clients sent data. The second layer of sharding was a consistent hash ring setup that could be grown pretty much boundlessly. I suspect that same project might still be going, and, unless they changed other parameters, could easily have 10k+ tables now.

                      1. 1

                        I’d be interested to hear someone from Facebook chime in on their mysql usage.