Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 1 year ago.
This post was edited and submitted for review 1 year ago and failed to reopen the post:
Original close reason(s) were not resolved
Improve this question
My app involves repeated equality comparisons of many megabytes of an array of strings. I was (naively?) surprised that such comparisons became immensely faster when I added a shadow array of the count of each string; if the counts are not equal, the strings cannot be equal.
I would have thought that the Swift compiler could easily and efficiently maintain a string's count in the internal representation, and short-circuit string equality comparisons if the counts differ. The initial creation of each string would provide the opportunity to initialize the count, and each manipulation would then know the count of each participant. But it seems not so. If only some applications would benefit from a stored count, a compiler switch to turn it on could help them.
Would this be a reasonable standard library refinement?
The problem with String is that it is not stored always in the same way. There are different types of storage and for them to be properly compared, they have to be converted to the same representation first.
Looking into the source code here: StringComparison.swift, I think you are right and count should be compared first when checking equality.
I recommend filing a bug.
Related
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
This question does not appear to be about programming within the scope defined in the help center.
Closed 3 months ago.
Improve this question
I have been looking around in the documentation, but could not find an explanation of what the "int2vector" type really is. It is found in some system tables like pg_trigger, but that's all a documentation search returns...
It seems to be vaguely similar to int2array, but has a different OID (INT2VECTOROID is 22, INT2ARRAYOID is 1005).
I have found ways to generate int2array in SQL (for instance with SELECT cast('{1,2}' as int2[])), but not int2vector.
The question applies to int4vector & int4array as well, and the uses case is when interfacing with libpq in binary format.
int2vector is an obsolete data type for arrays of smallint from the time before PostgreSQL had array data types. Today you would use smallint[].
You shouldn't use int2vector in your table definitions, as it is not documented. On the other hand, it is unlikely to get removed, since the catalog tables make use of that data type for historical reasons. There is no advantage to be had by using int2vector.
Closed. This question needs debugging details. It is not currently accepting answers.
Edit the question to include desired behavior, a specific problem or error, and the shortest code necessary to reproduce the problem. This will help others answer the question.
Closed 1 year ago.
Improve this question
I have this code, and it is supposed to write to Firestore. However, when it does the function, in the back end, it shows highlighted as red and then disappears.
db.collection("jokes").document("Dad Jokes").setData(["\(dadJokeNum + 1)": Joke.text!])
Please help.
Using SetData without Merge will delete existing values - it is suggested that if the document exists prior, you should always be using merge: true. Firestore also has a limited 1 write per second each, you should be managing writes together as it is most likely similar writes will conflict and potentially resolve out of order.
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 1 year ago.
Improve this question
I'm transitioning from a SQLite implementation to CoreData.
In SQLite, searches are fairly limited. In a typical search, for a string like "card", I would want to know if any members of a set of letters like [n,l,j,x], would be a valid part of a word, or a whole word, in a stored dictionary of strings.
So, in the example above, I would have to look for "nard","lard","jard","xard" and then repeat that process for each subsequent position in the string: "cnrd","clrd","cjrd","cxrd".
This is slightly controlled by the fact that I only need a single match per position in the target string to "qualify" it, so I can search for "nard","cnrd","cand","carn", and if I get a match at any point, I can mark that point in the target word as qualified, and focus on the other targets.
Thus, if I got a match at "nard" and no other matches, the next loop might check "clrd","cald","carl", and so on. If I got matches at "nard","cand", the next loop would be "clrd","carl" : you get the idea.
Does CoreData, which I know under the hood is just SQLite anyway, offer any more advanced features that would allow me to improve the default algorithms I've used, perhaps using regex? Can a pattern like \^{3}[nljx]\ be somehow used?
I'm not at the point where I'm writing the code to experiment in this direction, so anything people can point me to is great.
When you use a SQLite store with Core Data, predicates are translated into SQLite code and executed in SQLite. Predicates therefore have SQLite's limits on what's possible. Core Data can use other store types with different capabilities and limitations-- for example, you can use a predicate that's any arbitrary block of code, but the entire persistent store gets loaded into memory all the time. Whether one of those would work for you depends on how much data you have.
Yes, you can use a predicate with NSMatchesPredicateOperatorType to do regex searching. SQLite doesn't support regexes directly, but Core Data registers a custom NSCoreDataMatches function to do the work without bringing everything into memory.
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 4 years ago.
Improve this question
What's the best practice for getting the last added element (let's say we know that because of a created_at field on the resource).
Should it be a call to the get all with max results on 1 like:
GET ../rest/v1/article?page=0&size=1&order=created_at,desc
and will return an array of one element
or maybe an "special" call like:
GET ../rest/v1/article/last
and will return an element.
I am looking for a best practice if there's one pattern for this.
Thanks!
I'm not a RESTful expert, but in my opinion the first solution seems the best.
The second is more practical, but routes are often associated with resources, the addition of a "last", especially preceded by a "/" seems strange to me.
In addition, API users usually use the sorting parameters, and what about users which need 10 last elements ?
If you add something after ../rest/v1/article, it must be an ID for one particular element, a sub-resource, or for actions that are outside the CRUD like ../rest/v1/article/:id/subscription.
Both URLs are RESTful and identify a Resource. Of course the first would return a collection containing a single element while the second would return this element directly.
The first form will be automatically supported if you support paging and sorting at all.
You write
or maybe an "special" call
I don't see this as an 'or', it should be an 'and'. The second form is optional and it could be helpful. If you have an URL pattern like
GET ../rest/v1/article/{id}
it is easy to implement logic that can distinguish normal IDs like, for example, 123 or A567 from special IDs like 'last`.
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 6 years ago.
Improve this question
On one hand I want an ordered collection, on the other hand I want every item in the collection to appear only once.
I can either use an array and sort it every time I insert an item - and insert only if not in the array.
or use a Set data structure and sort it every time i query the data
Does someone have better solution?
There are several third-party libraries implementing an ordered set in Swift, so you could check them out.
Also, you could write your own implementation of an ordered set (you can base it on an existing one) if it is not an overkill for your task. The way you choose really depends on your app.
And in the end, you could use one of two ways that you proposed: using a built-in array or a set. In order to choose between them, take a look at your app: what action will be performed more often? Getting an access to elements in order (use array then) or addition/deletion of existing elements (probably, the set is the way to go).
This part was edited based on comments below
If you go for an array, note, that a built-in contains for arrays will not know that an array is sorted, so it will probably be O(N), not O(log(N)). So you should either write a custom replacement for the contains method, or (this is, once again probably a better way), write a custom collection class that implements contains the right way (however, since contains is a protocol extension method of SequenceType, my knowledge of Swift, I'm afraid, is not good enough yet to tell you how to do it properly, maybe someone else will).
UPDATE (based on your comment to your question):
I believe, in your particular case (a chat app) array is superior. You only have to sort old messages once, and you will not probably try to add very old messages once again, you only have to make sure you don't add new messages twice (it is implementation-dependent though, so you know better, I'm just assuming). So you only have to check that the last messages in your old array do not overlap with first messages in the array that you add. Sort of :)