so you can determine rank by self-cross joining, and checking that t1.x<=t2.x. But, if you want to gen rownumber and not rank, how would you go about doing that? is that mention disqualified or is there a trick that can be used?
No UNIQUE column required, but a unique ordering must be possible
This method requires no single unique column but it requires a unique order. You can find it this approach in SIMULATING ROW NUMBER IN POSTGRESQL PRE 8.4 by Leo Hsu and Regina Obe, it's listed as the "cross-platform method." That ...
so the only thing I don't have that I want, is a method to assign row numbers, 1) without a sequence, 2) without window functions, 3) without a unique ordering, 4) without any possible unique key -- on column, or composite on column.
It might be nice if suggested edits review indicated if the person suggesting the edit is the question asker. Also showing comments in that view might be handy.
@TypoCubeᵀᴹ you around? You commented on this earlier
Curious if we should close it or leave it open for now or ... It's a good question in that it's not likely to be easily googled, but it's a poor question in that it's just basic sql parsing.
I've reworded, I'll fix it up a bit more too. You guys are way off on the close votes, ffs. That's a really good question about SQL parsing. I'm interested in it.
Hi, I'm looking for a LaTeX package to draw e-r diagrams and I've found the tikz-er2 package (texample.net/tikz/examples/…), but it seems to be very basic. Is there any better?
@EvanCarroll you might want to reword it now that I've changed the question — also the bit about max(*) is a red herring, the alias stuff is redundant now (and trivially basic)
You still have a table without a FROM. It's the (imaginary, dummy) table with 1 row. That's what Postgres does (and SQL Server. and MySQL). That's why 1 is returned. Nothing really spectacular here. Mind that even these 3 dbms have different rules. Some allow SELECT .. WHERE .. without a FROM. Some do not (I won't spoil the surprise, you can check ;) — TypoCubeᵀᴹ13 mins ago
^^^ this is the most useful set of words on the page
Also, the "Speaking for PostgreSQL" perspective of your answer is why these badly scoped questions aren't a great fit — a good answer should really cover all the RDBMSs that allow the syntax (but is anyone really interested in that? idk)
There is a vim.stackexchange.com, a saleforce.stackexchange.com, an ubuntu.stackexchange.com (because unix.stackexchange.com was not good enough for those kids). etc.
@TypoCubeᵀᴹ no, you can't actually add rows to a table with 0 columns.
over the implicit dummy table that pg already has?
I would ask what is the use of an explicit Dummy table in PostgreSQL and let you answer it, but you'd probably downvote yourself for the uselessness of the question.
Can someone mitigate this to Meta Security? Just realised this is best placed there.
I have a program that hashes some passwords and saves them to .txt files, I'm not using databases for storage.
My questions is, is it considered more secure to save hashed passwords in one file or to have them ...
@EvanCarroll You still haven't answered those comments - under the question I mean. We do make these comments for a reason. So the OP explains why they want or ask something, and for clarification. When the OP (which was you in that case) does not respond to comments, I take it as being rude or very new to the site. Or has abandoned the question.
I can go backport the question from the real life example. I ended up using row_number. I just wanted to know how not to use it because you can't use row_number() in join conditions, or in alter table commands without doing a self-join.
@EvanCarroll You've got some good answers, and you seem to know your stuff concerning PostGIS — an area we can use more expertise, but you would learn faster here if you recognized that other users are much more knowledgeable than you on many other areas
@EvanCarroll and that should be in the question - that you may need/use this in updates or joins or whatever. That's why people were making comments. You asked question X while you wanted Y.
if you don't mind that, and feel you have learned from asking it, that's fine — but if you want your questions to be useful to others and to remain open, you are going t have to learn to work with the more experienced users here
@EvanCarroll that question is not of interest to anyone with expertise here
@MaxVernon I would countervote if I could - because it is a common DBA problem - identify a good technical PK for duplicate removal. And for huge tables some solutions dont work well — filiprem2 days ago
@EvanCarroll if you formulated your problem more generic way it would be more interesting. Why do you need monotonic increasing numbers from 1? Is it not enough to have unique row ID? personally I use ctid as last-resort sometimes... — filiprem2 days ago
I provide two methods of generating row numbers and answer them in painfully explicit detail. If he's interested in the performance implications he has a starting point that otherwise wouldn't exist at all if not for my contributions to dba.stackexchange.com
@DannyWatson well I haven't deleted or even downvoted Evan's answers — I engaged in trying to help him be a more constructive member of the site though, which basically boils down to "don't ignore feedback from other users"
Thought it was too cringeworthy that they CGI'd the late Tarkin, I won't spoil it for you Typo but I really enjoyed it, enjoyed it more than Force Awakens
@JackDouglas People forget that the help they receive is purely voluntary, sometimes it's best just to put feelings aside
I had a question it was pretty simple and clean cut. None of the comments were on clarity of the question at hand. I asked,
How can I generate a row_number without using a window function?
The question wasn't totally downvoted.. It had one upvote.
The question wasn't specific to just me. While...
What is the difference between: SELECT Count(*) FROM Table and Select Count(*) Table
Can someone mitigate this to Meta Security? Just realised this is best placed there.
I have a program that hashes some passwords and saves them to .txt files, I'm not using databases for storage.
My questions is, is it considered more secure to save hashed passwords in one file or to have them ...
