Comments by csrss (Top 58 by date)

Thanks for the explanation! I'll look into this later.

Is there a solution without using 'OR' maybe?

I am just considered about using 'OR' in this solution. As I have mentioned, OR is very slow.

Thanks! This helped, I had to adjust it to my needs of course but this is it :)

Thanks. Will try that out

Thanks

Thx :)

Thx!

Unfortunately, this does not work

So far, this is the most relevant answer. However, I did that already, and though that there is an easy way, like with model binding. I'll probably stick to my current solution. Thanks anyway

I dont really see how this can help

Well, I am not asking for an advice what route should I take, right?
Maybe creating such attributes is not possible and there are just built-ins.

Only for controller method

Anything, extraction of user role, auth token, any sort of information, required to run an endpoint.

The question is correct I think, because [FromBody] is an attribute, but it is some sort of built-in attribute. I am already doing it like it is show on the link you provided, like this:
[ModelBinder(typeof(MyModelBinder))] MyClass myClass
But it feels that it can be done better, something like this:
[FromRequest]MyClass myClass
And this is what I am asking about, how to do it.

But even if it is what I am looking for, how can I get HttpContext, how can I set output parameters, etc - there is no information in this link you posted. I dont see, how can I extract data from HttpRequest and pass it up as a Controller parameter.

But this is something completely different

Thanks. I'll have a read.

Yes, I have checked. The only thing which is odd to me, is:
1. In AspNetRoles there is a column Discriminator which is set to "Role"
2. In AspNetUserRoles there is a same column Discriminator which is set to "IdentityUserRole<string>".
Maybe because of this it is failing. Anyways, I have used different design where joining Collection<identityuserrole<string>> to user is sufficient, so I just leave it like this for now.

Nope, this does not work.

Yeah, if I update to ef core 5, a lot of queries stop working or are a lot slower, so this is why I stick to 3.1. This is strange, I have very similar thing like you do, then I do something like this:
db.Customers.Select(c => new
{
c.Name,
Currency = c.Currency ?? c.Firm.Currency,
}).Select(x => new { a = x.Currency.SomeListProperty.Where(...)})
And everything crashes. Everything is initialised with huge .Include().IncludeThen() etc

My setup is this: .NET 5.0, ef core 3.1.

First one does not work. Second one, cannot be done, because I need to run a query on selected property. So, my scenario is this: if client has a policy - take, if not - take company policy and then run a query for all rules contained within policy object, so this is .Where(...) on a result, returned by ternary operator and then there are next steps, and all this should be executed within one request to database, and all is Linq to sql. Client side execution is forbidden.

Yeah, I understand. Thanks anyway. The more I am trying to do with this linq to sql - it is just crap and nothing else. Cannot translate the simplest queries. It can just perform Where() couple of times, some select and that's it. Nothing beyond.

Yes, tried to do it that way, same result.

Yes, I understand it cannot be translated, and that it runs on server, but there should be some way to do it like this?

All objects are valid, non of them are nulls. But it seems to me that this line to sql is just dumb. Because even simple SelectMany fails, but there are no nulls.

Thanks

I actually made it work

I need expressions, because I am fetching 100% dynamic data.

Yes, but I know about finding most common numbers in int array. The question is about algorithm to cluster data probably. I just wanted to know, if there are any applicable.

Yes, the nearness should be calculated based on numbers in array. So we can see, that 10 is near 9 and near 11, but not near 100, or 1. From what I know, there are some clustering algorithms out there, but not sure if they will fit here?
Language is C#

Yeah, probably hard to explain over the net. There is your regex:
^(?P<text>\w.*)(?P<num1>[\d]{2,},[\d]{2}|)\s(?P<fromdate>\d{1,}\s\w{3})\s-\s(?P<todate>\d{1,}\s\w{3})\s(?P<year>\d{4})\s(?P<num2>\d{2,},\d{2})$

And those are my 2 lines:
My super data plan 10gb mobile 99,00 1 jul - 31 jul 2020 99,00
My super data plan 10gb mobile 1 jul - 31 jul 2020 99,00

When I put everything in regex101 website, on the right side, I can see the matches. If I have those 2 lines, I will have 2 matches with the above regex, however <num1> will not be captured. If I remove 'or' from <num1> group, it will match first line and capture <num1> but will not match 2nd line. I am trying to match both lines, and capture <num1> in a first case, but omit it in 2nd case, so lets say I would have an empty group or no group - I have no idea how it works with regexes because I have little to no knowledge about regexes.

Thanks, is it possible to capture <num1> in first case, while matching both lines?

For now I am working in regex101 website, PCRE (PHP) is selected as Flavor

Sorry, maybe I expressed myself in a wrong way. Lets say, there are 2 lines:
My super data plan 10gb mobile 99,00 1 jul - 31 jul 2020 99,00
My super data plan 10gb mobile 1 jul - 31 jul 2020 99,00
I am trying to understand how to match both of them

And if you make it like this:
^(?P<text>\w.*)(?:[\d]{2,},[\d]{2}\s+)?(?P<fromdate>\d{1,}\s\w{3})\s-\s(?P<todate>\d{1,}\s\w{3})\s(?P<year>\d{4})\s(?P<num2>\d{2,},\d{2})$

Then 99,00 will be part of <text>

Nope, doesn't work. If you use this regex and remove first 99,00 - nothing will be matched.

Yes, but in this case, you have removed <num1> from regex

Hmmm, I have modified your regex this way:
^(?P<text>\w.*)\s+(?:(?P<num1>[\d]{2,},[\d]{2}))?\s(?P<fromdate>\d{1,}\s\w{3})\s-\s(?P<todate>\d{1,}\s\w{3})\s(?P<year>\d{4})\s(?P<num2>\d{2,},\d{2})$
Because I want to make a <num1> optional. But so far I have no luck in making it work, because if I remove <num1>, then a line is required 2 spaces. If I make one space optional - this is not working. Any idea, how can <num1> can be made optional?

Yep, thanks. Still space is needed of course:
^(?P<text>\w.*)\s+(?P<num1>[\d]{2,},[\d]{2})\s(?P<fromdate>\d{1,}\s\w{3})\s-\s(?P<todate>\d{1,}\s\w{3})\s(?P<year>\d{4})\s(?P<num2>\d{2,},\d{2})$

First of, thank you. It works.
This is strange, I am testing regex on regex101 website, while this works when I open it on one computer I am now, does not when opened on another laptop - on which I was testing. Magic.

I have updated my question

I have add my sample data. I am trying to do recurrent grouping.

Maciej, I did the following way:
public static Dictionary<tkey1, dictionary<tkey2,="" tvalue="">> Pivot<tsource, tkey1,="" tkey2,="" tvalue="">(this IEnumerable<tsource> source, Func<tsource, tkey1=""> key1Selector, Func<tsource, tkey2=""> key2Selector, Func<ienumerable<tsource>, TValue> aggregate)
        {
            return source.GroupBy(key1Selector).Select(x => new
            {
                X = x.Key,
                Y = x.GroupBy(key2Selector).Select(
                z => new
                {
                    Z = z.Key,
                    V = aggregate(z)
                }).ToDictionary(e => e.Z, o => o.V)
            }).ToDictionary(e => e.X, o => o.Y);
        }
This one I took from the net. So after applying the above, next thing I did is:
.Pivot(x => x.Selector1, x => x.Selector2, (x) => x.ToList())
.Select(x =>
{
var obj = new ExpandoObject() as IDictionary<string, object="">;
obj.Add("Column1", x.Key); // mandatory known column
foreach (var val in x.Value)
{
// dynamic columns
obj.Add(val.Value.First().FieldName, val.Key);
}
return obj;
})

However, this is not sortable solution, because dynamic columns do not exist in same manner as known columns in dynamic object, so I just cannot sort with LINQ, of group, etcetera.
So I decided to stick with T-SQL and dynamic queries + stored procedure.

You see, I have already tried such solution as yours, but unfortunately it does not work.

There is no numeric data - all rows and columns consist of string values. No, I cannot do it like you say - pivoted table is applied to a bigger query so the result table is returned which has all possibilities of sorting and filtering data data.

Thanks but no. I mean, neuro evolution does not learn "on the fly" as it name may suggests. Evolutionary network evolve in a context of finite data. So, in fact, it acts in the very same way as, lets say, basic MLP. The only difference here is topology. So any combination with "regular" learning is pointless. Well, it seems that neural networks are limited to handle finite data. Thats it. NN is not much more then a regular algorithm.

Thanks for the bayesian keyword - I'll check it up. Still, if it cannot be trained "on the fly" - this is not a solution.

Richard, thank you a lot :)
I actually changed my design just a little bit by introducing additional non-generic interface and it solved the problem.

So it seems then I cannot use current design, because interface (ITest) uses T as in and out both. Thanks

Here is formatted json:
[
[
"test",
{
"prop1": 1,
"prop2": "test",
"prop3": "test"
}
],
[
"test2",
{
"prop1": 2,
"prop2": "test2",
"prop3": "test2"
}
]
]

I still got no idea, what the object structure is.