Same here. We are moving from mongo to postgresql. And we are going to rewrite our geolocation services from C# and mongo to D(vibe.d) and PostgreSQL
Dne 31. 8. 2017 7:15 odpoledne napsal uživatel "Matthias Klumpp via Digitalmars-d-announce" <[email protected]>: On Thursday, 31 August 2017 at 11:56:53 UTC, aberba wrote: > On Wednesday, 30 August 2017 at 16:53:40 UTC, Matthias Klumpp wrote: > >> On Wednesday, 30 August 2017 at 07:47:53 UTC, Sönke Ludwig wrote: >> >>> Apart from removing the old vibe-d:diet package in favor of diet-ng, >>> this release most notably contains a number of performance improvements in >>> the HTTP server, as well as improvements and fixes in the WebSocket code. >>> Furthermore, initial OpenSSL 1.1.x support has been added and a few @safe >>> related issues introduced in 0.8.0 have been fixed. >>> >>> Change log: >>> https://vibed.org/blog/posts/vibe-release-0.8.1 >>> >>> DUB package: >>> https://code.dlang.org/packages/vibe-d/0.8.1 >>> >> >> Debian packages are on their way too :-) (pending approval from our >> archive masters). Granted, this is most useful for Vibe.d-using software >> that wants to be in Debian. >> >> Now, the only thing I am missing in Vibe is a good interface to >> PostgreSQL, because in some circumstances MongoDB is just a very bad choice. >> (Postgres even outperforms it in my testcase, and PG supports JSON/BSON >> as well now). >> DPQ2[1] looks very promising though :-) >> >> Thank you for making Vibe.d! >> >> [1]: https://github.com/denizzzka/dpq2 >> > > Even with mysql (using mysql-native), the absent of something like > > struct User { > @optional int userName; //its ok if row doesn't have this column > @as("phone_number") string phoneNumber; > .... > } > > User[] users; > .... > foreach(row; ...) > { > users ~= row.toStruct!User; > } > .... > I would love that :-) MongoDB has facilities for these stuff automatically thats why using it > seem convenient and speeds up stuff. > Yeah, it speeds up development, but that doesn't help much if Mongo swallows your data or messes up replication, or if your data by its nature simply does not fit a document store well. I've found Postgres to be very reliable and also very fast in the past, and - in large part thanks to Vibe.d - I am using Mongo now in a test deployment instead of it. I don't feel comfortable at all in continuing to use it though, which is unfortunate, since the development speed with it is quite high.
