Thanks!

I am glad you are finding my content valuable. Thanks for the note. I will try to pay attention to that.

I am glad it was helpful. Caching itself is actually fairly easy. The difficulty will come in when it comes to deciding when to invalidate the cache and get fresh data.

@@IAmTimCorey Exactly. Some people suggest to Dispose and create a new fresh instance.. but I don't think it's a valid solution.

Thanks!

You are most welcome.

Yep, it was already in progress.

Glad to hear that!

I am glad you are enjoying them.

Yeah, I'm trying to figure out what the issue is.

@@IAmTimCorey Sounds almost like a directional mic that was pointed the wrong way. Picks up your voice indirectly so it is missing a lot of signal.

You are welcome.

Excellent!

I'm glad it came in time.

You are welcome.

You are welcome.

You are welcome.

You are welcome.

You are welcome. I am glad it was helpful.

You are welcome.

You are most welcome. Thanks for watching.

You are welcome.

You are welcome.

Yep, that's true.

Not in the short term, but that is something I am considering.

You're very welcome!

Thank you!

You are welcome.

Thanks!

When we build desktop apps, we tend to get a bit lazy with our data because we have such power available to us on the client side. If you think about 1GB of data, and how much content that represents, you will see that it is too much to display to the user. Yes, you might "need" all of that data but the user never sees that much. If we converted 1GB into 200-page books, we would need over 4,000 of them. So when you move to the web (or honestly even before since the same principle is true on the desktop - the web just forces you to work differently), you need to think about what is actually needed. For instance, if you have a list of employees, you probably just need a few pieces of information (name, Id, department, title), not all of the information about an employee. So, you create a query that gets you just that information. Then, you don't need to display all of the employees for your entire organization on the screen at once (even hidden behind a scroll). Load just the records they need to see. That might require you to do multiple queries if they want to scroll manually through the list. You can use the Virtualize panel in Blazor to help make that more transparent. You will have to do some work, but that work should be done in both apps, not just the web app.

@@IAmTimCorey I see this almost obsessive trait, of wanting to move from desktop to a web environment and it is just crazy to me. In the end, it has little thought for the end user and thinks only about the developer. The user experience will be objectively and measurably worse in a web application, there is no denying this, you even hint at that yourself, one of the many, many reasons. Don't move to web because some manager who does not care the slightest about the end user experience wants to use web-devy stuff. Think about the end user and the problem you are trying to solve, instead.

You are welcome.

You are welcome.

You are welcome.

@IAmTimCorey hey, mr Corey.. I am a South African that cannot afford your full courses... How's about a give away video? Hmmm 😁😁

have you found out how to achieve that?)

Time isn't really the issue here, it is the content. So what content are you hoping to cut? These videos are about preparing you to use a topic in the real world, not just a "here is what this topic is"-type. I do have the 10-minute training series for those types of videos.

Thank you!

Yes, you can clear the values. Just use the Remove method: docs.microsoft.com/en-us/dotnet/api/system.runtime.caching.memorycache.remove?view=dotnet-plat-ext-6.0

You are welcome.

You are welcome.

You are welcome!

Yes, it is thread safe.

locks? Can't think of anything else

Good question. It depends on if this is a problem. First, this probably won't happen often. However, if it does happen, the worst-case scenario is that you call the database twice instead of once and that you overwrite the cache with the second call to set the value. But if that is a serious problem, you could implement a service that does caching in order. Then it could deduplicate requests. I think that's a lot of extra overhead for every call when the problem it is solving has such a low impact.

@@IAmTimCorey I was also wondering about the thread-safety of memory cache... although, tbh, my main concern is about thread-safety of Redis cache, as I intend to use memory cache only in dev environment, and Redis in production.

Thank you!

Different room and I’m still working through the best setup.

Typically the best option is to invalidate the cache and reload it.

I don't. I don't really cover EF, except for one video.

It is in the memory of the computer running the application. So for most web projects, that means the server. The exception would be for Blazor WebAssembly projects, where it would be in the client's memory.

The top two results will cover almost all of your threading needs: kzfaq.infosearch?query=async

@@IAmTimCorey thank you sir for response

Here you go: kzfaq.info/get/bejne/o9Wida6i2dfRo6M.html

@@IAmTimCorey thank you again 🙏

Thanks!

HELLO..did you watch the video? he showed it. In method you calling cached data, you should populate it if cache is expired/null. So just read data and put it in cache. so you got auto repupulation. If you want auto populate cache even if nobody ask for data - then your idea is dumb and you should not develop apps, because it's dumb to keep data in memory if nobody even asks for it.

@@SayWhaaaaaaaaaaaaaaaaaaaaaaat Hi Tomas, I did watch the video but I would like to repopulate the data after it expires automatically. My ASP Core app has queries that need 15+ seconds and I cant ask the user to deal with that or do it myself each morning. I have seen examples with PostEvicitionCallbacks and ExpirationTokens that should solve the need but I am having trouble.

In RAM, although if the memory pressure on your web server, your web server might offload some to disk. However, that is a transparent process to you.

It is C# in general. I talked at the end about how you would use it in a desktop app.

Yep. You will just need to manage your cache invalidation periods so you don't cause an issue.

Worst case, you double-set it. Not a big deal, since if you didn't have caching then you would be getting that data every time. If your data would change based upon the new call then you shouldn't be caching the data.

Store the data the user will actually need. For instance, do they actually need a million records? If so, caching probably needs to happen on Redis or something similar. Another test is what are they coming back to repeatedly?

Watch the Suggestion Site App course here on KZfaq. We will be using in-memory caching in the app.

@@IAmTimCorey Ok got it. Thanks

Yes, between the time you check for records and when you set the records in the cache, someone else can do the same thing. That would result in multiple people doing the data lookup from the database. That's not a huge deal in most cases. If it is, you can use LazyCache to wrap this so that you only get one database read.

That would be a lot of work for little return. You could create a class with a completion event and an indicator when the class has been activated. Then, the first call activates the class and subscribes to the completed event. Every subsequent call could subscribe to the event only, knowing the call is in progress. However, that's a lot of additional overhead when it usually isn't necessary. Caching isn't about never making a duplicate call. It is about reducing the number of calls. You should be able to function without caching. In the case of your 10-second call, the first thing I would do is work on reducing that number. 10 seconds is a LONG time to wait for something. If you could add some complexity to reduce that call to 3 second, the number of duplicate calls you would have would go down and the overall experience for everyone would be better. Another option for long-running calls like that is to only let the system kick them off. Always work from the cache and have the system handle refreshing it in the background. That way, you never have duplicate calls, and the initial call happens on system load, so no one is ever waiting those 10 seconds.

Memory pressure. If your memory pressure gets too high, you can flush the cache. That's not as easy with a singleton.

I believe so. Just remember that you are storing it on the client's machine so it will be client-specific.

In memory is quick and for small jobs. Redis is for when those small jobs get larger.

On the server unless you are running Blazor WebAssembly, which runs on the client's machine.

No, this is outside the scope of the C# Mastercourse.

I will add it to the list. Thanks for the suggestion.

Sure, just set the cache in that event and it will overwrite the current list.

@@IAmTimCorey Thank you so much

The cache is thread-safe.

> because every hour ... I will reload You can write... but do you need to write something? All already written and need just set 1 hour timeout. > if there are something added You can update your cache any time you want. It just hashtable in the memory - use the same key and replace the object.

@@AMurr-zm4wh if we set the Memory Expiry to 1 hour then after 1 hour we need to reload again (IO cost) but if I check just count of records every hour maybe I don't need to reload again for a couple of hours and it means less IO cost (I need to add new records to my pool only)

@@nadervaghari51 Ok. Set 2 hour timeout and update cache in 1 hour. Take object out of the cache, add if you have something to add to the object and put it back into cache - this will reset the timer. In this way you not need to write any additional code.

@@AMurr-zm4wh but the main question was can we use Singleton fields in place of In Memory Cache or not?

@@nadervaghari51 Yes, you can. At least I use it in the past in global.asax - there was no other way at time. But main question is different - Why do you want to write and support additional code in place where nothing need to be written?

You can always turn on CC in KZfaq

The course isn't about in-memory caching. The course is about the API project template. We cover a LOT of ground in that course, but we don't go into depth on things beyond the scope of the course. So yes, you will find more in-depth topics on this channel or in other courses that are more specific to the topic you are looking into. The API course goes into depth on API, not on all of the connected items. Otherwise, it would be a MASSIVE course. Also, the primary purpose of a course is to lead you down the path of learning a technology in a structured, ordered manner. That is the main benefit of the course. There will always be more depth to go into, more things to learn, and more areas to grow in after you gain the foundational knowledge of the topic.