# JuliaDB error loading tables

**URL:** <https://discourse.julialang.org/t/juliadb-error-loading-tables/47795>\
**Category:** Data\
**Tags:** question\
**Created:** [October 5, 2020, 3:51pm UTC](https://discourse.julialang.org/t/juliadb-error-loading-tables/47795 "2020-10-05T15:51:48Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![David\_Ortley](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/david_ortley/32/7845_2.png) [@David\_Ortley](https://discourse.julialang.org/u/David_Ortley)\
**Post date:** [October 5, 2020, 3:51pm UTC](https://discourse.julialang.org/t/juliadb-error-loading-tables/47795/1 "2020-10-05T15:51:48Z")

</div>

I frequently have problems reloading the JuliaDB tables that I save. Sometimes they’re smaller tables on the order of a few MBs and sometimes a couple of GBs, but never anything close to how much RAM my machine has. I’m currently looking at a dataset that generates an EXCEPTION\_ACCESS\_VIOLATION on Windows and crashed Julia, but sometimes loading a table results in a stack trace, it’s mixed.

Two questions:

Sometimes an older file I save that was loadable before isn’t loadable later on with an updated version of JuliaDB. Is the file format stable? I will admit that I haven’t been paying attention to which version of JuliaDB a file is being saved or loaded under, but I know that I’ve seen this happen after a package update updated JuliaDB. However, this problem might be remotely connected to my other question below.

I may be using JuliaDB incorrectly. I work with a lot of time history data (time vs pressure, for example), and have been trying to store the time and pressure arrays along with associated meta data in a single table. So a row might look like `JuliaDB.table((event_date=[timestamp], event_t = [time_array], event_p = [pressure_array]))`

Where `time_array` and `pressure_array` could be empty `Float64[]`'s or could be stuffed with actual times and pressures.

I know from a traditional database perspective, this is a no no. Should I not be doing this with JuliaDBs as well?

-David

---

<div class="post-metadata">

**Author:** ![joshday](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/joshday/32/368_2.png) [@joshday](https://discourse.julialang.org/u/joshday)\
**Post date:** [October 5, 2020, 4:37pm UTC](https://discourse.julialang.org/t/juliadb-error-loading-tables/47795/2 "2020-10-05T16:37:32Z")

</div>

> [@David\_Ortley](#):
>
> Is the file format stable?

The format is based on Julia’s serializer, so anytime a JuliaDB struct changes, that table won’t be loadable in a new version. That being said, I don’t think anyone has been touching JuliaDB internals for a while.

> [@David\_Ortley](#):
>
> Should I not be doing this with JuliaDBs as well?

A field with Array elements shouldn’t be a problem for saving, although yes, things like that can make data analysis a mess.