Brewery page

Could it be picking collabs from other brewery pages? Or aliases?

1 Like

Whose interests does it serve for the page to be up? For whom does this count as progress?

I just added a beer, rated it and refreshed the brewer page. As it loaded it first said I have rated 11 beers from the brewery, while elsewhere on the page said Iā€™d rated 14. Then it finished loading and said Iā€™d rated 28 of the breweryā€™s 24 beers.

2 Likes

Only serves the interest of makeing people move to Uā€¦

2 Likes

Highly doubtful. For example it showed me that I have rated double the beers than a brewery actually had - and there are no aliases there and they didnā€™t have that many collabs with others, nowhere near.

1 Like

All of the ones I have noticed have told me precisely double the number of beers I have actually rated, which implies it is just a coding error where it accidentally counts everything twice.

2 Likes

Exactly my understanding as well, so if this is an easy fix, this will not get resolved this year, far more importants data to screw up firstā€¦:smiling_imp:

5 Likes

Ok. Like that i can sort on my score, Love the light colouring so its cla the ones i had.

Love that ctrl-F now seems to work

New frustration is i sort, click on beer, go back and then i need sort again. its lost the order i wanted
But WAY WAY better than last week

3 Likes

Also I have just spotted the small print above the beer list telling me how many beers Iā€™ve rated from a brewery.

Just to know. When I type letters on the Search menu of the beer list, sometimes, some letters are simply jumped because of the lag and the search wonā€™t return anything because of that. Am I the only in this situation? Am I typing that fast?

2 Likes

I do not know if anybody already submitted this feedback, but with the old page it was possible to see the last ratings received by a brewery. Now not anymore.

Would it be possible to have this feature back?

Thanks @joet @services

The Recent Ratings section was there on launch but went missing during an updateā€¦donā€™t know what happened with itā€¦I know it was hard to load and made the page lagā€¦maybe they removed it until they fix itā€¦

@aww @services ?

Do you mean this page?

no, it was on the brewery page itself.

It was this one if Iā€™m correct.

We have decided to remove it for the moment as it was contributing to excessive load issues. Itā€™s a really resource heavy process which involves pulling back all the reviews across the entire product range and then presenting the most recent ones.

The devs are still seeing whether we can achieve this in a more efficient way.

It was there before (in a different tab in the Brewer page) and it didnā€™t have any load issue at all. What causes those load issues now?

1 Like

SELECT reviews.*, users.name, beers.name
FROM reviews
INNER JOIN beers ON reviews.beerid=beers.id
INNER JOIN users ON reviews.userid=users.id
WHERE beers.brewerid=?
ORDER BY reviews.date DESC, user.id ASC
LIMIT 1

You do have indexes for all the appropriate key fields, I trust? (Actually, given what youā€™ve said, I donā€™t trustā€¦)

To whom should I send the bill?

3 Likes

We donā€™t run MySQL but why would you do a costly secondary sort by UserID if you arenā€™t using it? And any table.* select is needlessly costly if you arenā€™t using all the columns, and we definitely arenā€™t. Limiting to 1? Why? And whereā€™s your pagination?

So come onā€¦ letā€™s stop badgering the engineers with know-it-all advice. Itā€™s just mean.

The items are being looked at and will be gotten to in time. I do thank you for keeping these important items on my radar. We do have a heavy backlog but when it comes time to plan for the next quarter, itā€™s good to know what was left behind and continues to be irksome.

2 Likes

why would you do a costly secondary sort by UserID if you arenā€™t using it?

Iā€™ll address that later.

And any table.* select is needlessly costly if you arenā€™t using all the columns

HOW THE FUCK DO YOU EXPECT ME TO KNOW WHAT COLUMNS YOU HAVE!?!?!?

Sorry, mind-reading is not one of my skills. Why not criticise me for getting the table names wrong as well? (And the obvious typo!)

Limiting to 1? Why?

Because I have no idea how many were being displayed, Iā€™ve never seen the lists, as theyā€™ve never been of interest to me. Itā€™s an arbitrary number. 1 works well enough to satisfy ā€œthe most recentā€ and it should be obvious how to generalise 1 to any other arbitrary number.

And whereā€™s your pagination?

Pagination was not a feature that was even mentioned in passing. ā€œThe most recent Xā€ is just ā€œthe most recent Xā€, not ā€œa paged list of X in recency orderā€. Just coding to the usersā€™ requirements, nothing more. (Shocking, I know!)

However, I did put the capability for pagination in there.
And you noticed it.
It was the user id.
If you are doing pagination, you need to be able to uniquely identify the record to resume from and that must be unambiguously defined by the sort order. Two reviews could enter the database with identical timestamps, and therefore you need to be able to disambiguate them. I was presuming - not knowing what other fields might be available (remember, Iā€™ve not seen your schema) - that a user would never race himself, and thus his id would assist the timestamp for achieving uniqueness. Those with better internal knowledge should be able to pick the most appropriate field obviously.

So, back to the ealier question - ā€œwhy would you do a costly secondary sort by UserID if you arenā€™t using it?ā€ - well, Iā€™ve shown that you do need, if not that, then something to separate identical timestamps. However, on to the ā€œcostlyā€ aspect - the only time a comparison on that second field is actually performed is if there is a clash of timestamps, which should be very rare (you apparently think never). Itā€™s either essential, or itā€™s essentially free. Strangely youā€™ve taken a ā€œnot neededā€ and ā€œexpensiveā€ line. Maybe you need a better DBE.

So come onā€¦ letā€™s stop badgering the engineers with know-it-all advice. Itā€™s just mean.

Youā€™re employing the wrong staff, if they consider it mean. Engineers I know (been in the game 3 decades) have never taken offence at hypothetical solutions to know issues, they donā€™t take it as a threat. (However, Iā€™ve worked in nothing but open source software for over a decade, perhaps weā€™re a different subspecies with different attitudes to information sharing.)

1 Like