Added bulk update with single query

[thisago]

For now, just SQLite module is updating with single query, soon I'll implement in Postgres

Now I added bulk update for SQLite and PostgreSQL, The tests are running well

Closes #188

[thisago]

I think that I fixed all the requested changes in the last commit:

• moved changelog change to next release instead of making another version
• added comments to new funcs and types
• moved new funcs and types to outer scope
• renamed v counter var to argNum
• now the i var starts in 0 (not -1 anymore)
• added missing docs to new definitions
• removed commented code

[thisago]

I guess it's a WIP for now, amirite?

I'm already using it and looks good for me, the requested changes was done

[thisago]

but if you'd be interested to adding benchmarks that illustrate the performance boost that would be great.

haha, ok, there's a better way to do the benchmark or just run it with time?

[moigagoo]

but if you'd be interested to adding benchmarks that illustrate the performance boost that would be great.

haha, ok, there's a better way to do the benchmark or just run it with time?

I was planning to use benchy for it.

The way I see it:

1. Create benchmarks folder.
2. Add benchy to norm.nimble.
3. Add bench task to norm.nimble.
4. Add a benchmark that updates 1000 rows

[thisago]

So, the old implementation was replaced, I will put a copy of the old implementation in the benchmark file just for it, ok?

[thisago]

4. Add a benchmark that updates 1000 rows

1k of rows is too much to make multiple tests like benchy does... I'll limit to just one test ok? or you prefer to reduce rows, like 100 to repeat 10 times?

[thisago]

I ran 2 times the benchmark with 1k of rows and the result is very satisfactory: (for sqlite)

1th

min time      avg time       std dv  runs  name
   133.105 ms    133.105 ms  ±0.000  x1    Bulk update
139891.469 ms 139891.469 ms  ±0.000  x1    Update each row

2th

min time      avg time       std dv  runs  name
   129.890 ms    129.890 ms  ±0.000  x1    Bulk update
138162.341 ms 138162.341 ms  ±0.000  x1    Update each row

[thisago]

I updated the benchmark code to generate the DB just one time and copy into another file to modify.
I added also an indicator of DB generation progress, it shows a dot for every added row.

Also, an important thing to add next is bulk insert because inserting 1k of simple rows are too slow...

Another benchmark result:

min time      avg time       std dv  runs name
   182.585 ms    182.585 ms  ±0.000  x1   Bulk update
174604.512 ms 174604.512 ms  ±0.000  x1   Update each row

[moigagoo]

I'll do some minor fixes myself later and cut a new release.

Thanks for the great work!