OK. That makes sense. I thought those operations were "inherently atomic",
but they are really just doing an update like any other. The commit/retry
is fine, I was just trying to make sure I understood.
On Thu, Nov 16, 2017 at 2:17 AM, Luigi Dell'Aquila <
luigi.dellaqu...@gmail.com> wrote:
> Hi
Looks like a bug; report opened:
https://github.com/orientechnologies/orientjs/issues/285
On Mon, Nov 13, 2017 at 2:20 AM, Luigi Dell'Aquila <
luigi.dellaqu...@gmail.com> wrote:
> Hi Eric,
>
> The correct syntax should be just
>
> UPDATE x ADD list = {"ddd":444, "dd":44}
>
> If it doesn't work,
Thanks for the quick response. I'm not familiar with that option. What's
the syntax of the create property statement to do this?
On Thu, Oct 26, 2017 at 8:25 AM, Luigi Dell'Aquila <
luigi.dellaqu...@gmail.com> wrote:
> Hi Eric,
>
> Try to define the property as EMBEDDEDMAP with LINKEDTYPE =
h the old
> LOCAL storage engine.
>
> Anyway, even if implementing this should be quite trivial, Andrey could
> have a better idea about how to do that.
>
> Best Regards,
>
> Luca Garulli
> Founder & CEO
> OrientDB LTD <http://orientdb.com/>
>
>
> On 2
@Luca: How does it help that it used to exist in LOCAL storage? What I'm
looking for is a class in PLOCAL that works in every way the same as any
other class (specifically including replication), but is guaranteed to be
100% in memory and has the option of being persisted to disk or not.
--Eric
@scott: No problem--thanks!
On Wed, Jul 20, 2016 at 9:41 AM, 'scott molinari' via OrientDB <
orient-database@googlegroups.com> wrote:
> Hey Eric. Already done!
>
> https://github.com/orientechnologies/orientdb/issues/6447
>
> Hope you don't mind.
>
> Scott
>
> --
>
> ---
> You received this
OK. I'll do that later today. If anyone has any specific suggestions you'd
like to see as part of this request, please reply to this thread in the
mean time (and please label as "critical", "important", and "nice to have").
--Eric
On Wed, Jul 20, 2016 at 3:09 AM, Andrey Lomakin
Thanks Roberto. This isn't a current need for us (2D point-mapping is fine
for now), but it is important for our database architecture. On that point
(no pun intended), do you see any reason why OPoint and related Geo objects
wouldn't just be "extended" to handle 3D, 4D, nD points in the future by
OK. So the space used is just the ID itself? How many bytes is the ID?
On Wed, Jun 8, 2016 at 1:23 AM, Andrey Lomakin
wrote:
> Hi,
> >The second part is new--whether there is a difference between not adding
> a property and setting it to NULL
> Yes, there is the
Ugh and sigh. I had considered that (and tried it for a short time), but
ultimately decided I didn't have the bandwidth to go through that learning
curve.
On Tue, Jun 7, 2016 at 2:04 AM, 'scott molinari' via OrientDB <
orient-database@googlegroups.com> wrote:
> Yeah, I see what you mean. I
Yes, but unless I'm totally missing it, neither the old or the new
documentation really document the custom JS function environment. There are
some code examples that use some of the objects and functions, but, for
example (from Alessandro's example code), getId() and commit() are not
documented
Thank you. Very interesting. The first link is very useful, but it was
apparently removed from the current documentation. Wonder why...
The second link is one I've seen, but it's more high-level, focused on the
process of creating functions in Studio.
On Mon, Jun 6, 2016 at 10:52 AM, 'scott
This is very helpful, but how do you know the properties/functions
available on 'orient', 'g', and 'b' (i.e. command, commit, getId, etc.)? I
can't find this documented anywhere, short of a few examples here and there
(some of which conflict with each other)? Also, I'm assuming that this is
the
The first part, yes (I'm just stating my understanding that properties
defined in the schema that are not added to a record incur no overhead).
The second part is new--whether there is a difference between not adding a
property and setting it to NULL. In MSSQL, for example, there is no
difference,
That's too bad. If anyone from ODB is reading this thread, please make
documentation a higher priority! This is too complex of a product to learn
entirely by trial-and-error. In this particular case, my preference would
be to use custom JS functions, but after burning several days trying to
guess
Thank you, but actually, that doesn't get me what I'm looking for. True,
rid is a string in the JSON, but I'm still not able to get the rid as a
simple variable. What is need is just "#xx:yy", not part of a JSON
structure. In one attempt at a Javascript function, I used variations of
Another sub-question on this topic: I assume that if you don't add the
property at all, then it add zero bytes to the record--correct? But if you
add a fixed-size property (i.e. SHORT, INTERGER, LONG, etc.) to a record
with its value explicitly set to NULL, does that take up the same amount of
The time-series data is stored in a "cascade" of linked records (I'm
currently using LINKMAPs in documents, but the same could be done with
vertexes and edges); there are several possible structures, but one is
root->year->month->day->hour->minute->sample (this one allowing for
multiple samples
Ah so. Good catch. Yes, the position of the RETURN clause does appear to
matter. So making this change does eliminate the exception, but
unfortunately, it doesn't produce the desired result. In this case, no
record is returned *except *when the UPDATE succeeds. What I need is
essentially the
OK. That makes sense. The only reason I'm inclined to use Lightweight edges
is for performance reasons (is that not still a factor?). I don't see using
"promotable" edges (that start off Lightweight and may become Regular at
some point), but assuming there is still a performance advantage, I can
Ah so. It's a quirk when using IE (it works as expected on Chrome), but the
table display still shows [binary-data] for BYTE-type properties.
On Mon, May 23, 2016 at 9:26 AM, wrote:
> version 2.2
> browser chrome
>
> --
>
> ---
> You received this message because you
Thanks Scott. That makes sense. But it brings up another question that I've
had for some time. I've never actually created an ODB document database. As
I see it, there's really no reason to, since a graph database still retains
all of the features of a document database (at least the ones that I
Thanks for the details Luca. Just out of curiosity, are you planning to
phase out HZ over time or are you generally satisfied with this solution?
Of course the challenge with load-balancing in a distributed environment
(and probably moreso with a writeQuroum < totalNodes) is the trade-off
between
Yes, that's a good point. Since the OP's application was described as a
financial application, I assumed that these documents were created once and
essentially never deleted or "moved around", so I didn't figure there was
any need to worry about referential integrity. That said, using edges may
@Luca: Yes, that's understood. My question was not about what OrientDB *can
*do, it was really more about whether there were really any good reasons
to ever use properties on edges. So far, every use-case I've come across
where edge properties were a possible design solution, it seems like
2015-07-31 18:26 GMT+02:00 Eric Lenington e...@24x8.com:
@Luca: Yes, that's understood. My question was not about what OrientDB *can
*do, it was really more about whether there were really any good
reasons to ever use properties on edges. So far, every use-case I've come
across where edge
@Luigi: Excellent. Thanks for the quick and detailed response!
--Eric
On Wed, Jul 22, 2015 at 8:21 AM, Luigi Dell'Aquila
luigi.dellaqu...@gmail.com wrote:
Hi Eric,
at this stage the query executor is not smart enough to understand that
the two syntaxes have the same meaning, it just
@Alessandro: Ah, very nice! To me, this syntax actually makes more sense.
:) But regardless, it also seems more flexible. For example, I think this
should be possible, right?
SELECT
expand(out('haveMonth')[month=3].out('haveDay')[day=20].out('haveHour')[hour=10
and hour=20].out('haveLogs')) FROM
28 matches
Mail list logo