MySQL

Issue:

When adding a new dimension, we can't select the ne operator on the UI as shown below:

But we can add it in the json model like below (work around)

Issue split from hasura/graphql-engine#6951

our pytest output has a pretty low signal:noise ratio, and it's often difficult to understand the cause of a test failure, if there is one. [Example: a 8600-line log for what turned out to be a flaky test that passed on a second run](https://app.circleci.com/pipelines/github/hasura/graphql-engine-mono/10008/workflows/5e17c7

Working on a project I ran into the documentation for customizing the Foreign Key.
The docs read:

// https://sequelize.org/master/manual/assocs.html
// For example, to use UUID as the foreign key data type instead of the default (INTEGER), you can simply do:

const { DataTypes } = require("Sequelize");

Foo.hasOne(Bar, {
  foreignKey: {
    // name: 'myFooId'
    type: D

Sometimes it is needed to store compressed data in the DB. Unfortunately not all the DBs have built-in compression and FUSE compressed FSes are not available for every OS. So it may make sense to store compressed binary blobs in the DB.

Unfortunately when one sees them in DBeaver he sees them compressed, but often they are needed uncompressed. So it'd be nice to have a feature to decompress the

Bug description

On windows, I run prisma format and note the unusual file ending. The lines are all LF, but the very last line is CRLF.

This causes issue on my Linux CI where it formats it ending in LF's only, causing a diff to occur and the build to fail.

How to reproduce

1. On windows do prisma format
2. Open in HxD or similar
3. See attached:

![image](https://user-images.g

Currently we only support Upsert for CockroachDB. Implementing same for MySQL should be pretty straightforward as well.

Wonder if we can also use INSERT ON CONFLICT UPDATE for PostgreSQL to ensure operation with same signature also works.

Preflight Checklist

• I have completed all Troubleshooting Steps.
• I'm on the latest version of Directus.
• There's no other issue that already describes my problem.

Describe the Bug

I

Scenario:

• Single vttablet, connecting to external MySQL (happens to be RDS Aurora, but I doubt it matters).
• Single unsharded keyspace (keyspace3), no vschema, with a single table, containing a single row
• Test program using the Python pymysql driver, using the CLIENT.FOUND_ROWS flag, and updating a single row to it's current key value (i.e. no rows are updated, but a single row

Describe the bug
Using a time dimension on a runningTotal measure on Snowflake mixes quoted and unquoted columns in the query. This fails the query, because Snowflake has specific rules about quoted columns. Specifically:

• All unquoted column names are treated as upper case
• Quoted column names are case sensitive.

So "date_from" <> date_from

To Reproduce
Steps to reproduce

Issue description

This option will improve performance in many scenarios
https://dev.mysql.com/worklog/task/?id=8134

https://www.facebook.com/weixiang.zhai/posts/678596755543802

Consider the following table:

dsimple> create table t(val char(3));
dsimple> insert into t values ('dasdsa');
string is too large for column

Dolt should add to the error message the name of the offending column like mysql

mysql> insert into t values ('dasdsa');
ERROR 1406 (22001): Data too long for column 'val' at row 1