I see in the documentation of join the xrel and yrel. What does a relation mean here?
here rel(ation) means set of maps,
As I see it, it just states that in the context of clojure.set/join it's more meaningful to talk about relation instead constantly saying 'set of maps where the map-keys are the field-descriptors'
Related
I'm referring to :db/unique :db.unique/identity
(as opposed to :db.unique/value)
Upsert to me in my naivety sounds a bit scary, because if i try to insert a new record which has the same value (for a field declared unique) for an existing record, my feeling is that I want that to fail.
What I'd do if a field is declared unique is, when creating or updating another record, check whether that value is taken, and if it is give the user feedback that it's taken.
What am i missing about upsert and when/why is it useful?
A search revealed the following (not in relation to datomic)
"upsert helps avoid the creation of duplicate records and can save you time (if inserting a group) because you don't have to determine which records exist first".
What I dont' understand is that the datomic docs sometimes suggest using this, but I don't see why it's better in any way. Saving time at the cost of allowing collisions?
e.g., if I have a user registration system, I definitely do not want to "upsert" on the email the user submits. I can't think of a case when it would be useful - other than the one quoted above, if you had a large collection of things and you "didn't have time" to check whether they existed first.
Datomic property :db/unique can be used with :db.unique/identity or :db.unique/value
If it's used with :db.unique/identity, will upsert
If used with :db.unique/value, will conflict.
How can I retrieve an entity using the pull method by its entity id? I've used transact to add some datoms/facts (right phrasing?) to my db. I can see the entity id's if I do a simple query like:
[:find ?e
:where
[?e :arb/value]
]
The result being:
{[17592186045418] [17592186045420] [17592186045423]}
Now I'd like to retrieve one of these using pull. The examples in the docs for pull, however, use examples where the entity in question is associated with an id.
Specifically, the docs refer to an example from the musicbrainz sample data set, and the sample they suggest is:
(pull db '[*] led-zeppelin)
where (although the docs don't show this) led-zeppelin has been defined like so (as can be seen here:
(def led-zeppelin [:artist/gid #uuid "678d88b2-87b0-403b-b63d-5da7465aecc3"])
The docs say that the pull command takes three things: a db, a selector pattern determining I think what attributes are pulled for each entity, and the "eid" of the entity. So the above led-zeppelin var is somehow the eid.
I don't really follow totally what's going on there. The :artist/gid is a id attribute defined in the schema for musicbrainz it seems, and the third item looks like the the specific id. I'm not sure what #uuid is.
But, in my case, I have defined no id attribute for my entities. I was hoping to be able to use the unique entity id that I think is assigned by default to each entity. Is this possible? If so, how would this be done?
The solution here is simple. Just drop in the entity id number directly:
(d/pull db '[*] 17592186045418)
The mistake I'd made was to use the eid as a string, i.e. by double-quoting it.
Pulls third argument is a reference to an entity. You can either use one of the IDs that your query returned, or a lookup ref, like in the led-zepplin example, where you refer to an entity using a unique attribute value.
The purpose of a query is to find the EID of something given one or more of it's properties. If you already know the EID, you don't need a query, you just want to retrieve the attr/val pairs for that entity. So use the entity function:
(let [eid 12345
result (into {} (d/entity db eid)) ]
(println result))
Note that the result of (d/entity ...) is lazy and you need to force it into a clojure map to see all of the items.
Besides Datomic's own documentation, you can find more examples and unit tests in the Tupelo Datomic library. Besides containing many convenience functions, the James Bond example helps to clarify some of the Datomic concepts.
I'm surprised to find that query results in datomic are not lazy, when entities are.
Is there an obvious rationale for this choice that I am missing? It seems reasonable that someone might want to want to (map some-fn (take 100 query-result-containing-millions)), but this would force the evaluation of the entire set of entity-ids, no?
Is there a way to get a lazy seq (of entity-ids) directly back from the query, or do they always have to be loaded into memory first, with laziness only available through the entity?
You can use the datomic.api/datoms fn to get access to entities in a lazy way.
Note that you have to specify the index type when calling datoms and the types of indexes available to you depends on the type of the attribute that you're interested in. Eg the :avet index is only available if your attribute has :db/index set in the schema, and the :vaet index is only available if your attribute is of type :db.type/ref.
We use something like this at work (note: the attribute, ref-attr, must be of :db.type/ref for this to work):
(defn datoms-by-ref-value
"Returns a lazy seq of all the datoms in the database matching the
given reference attribute value."
[db ref-attr value]
(d/datoms db :vaet value ref-attr))
The datoms documentation is a bit sparse, but with some trial an error you can probably work out what you need. There's a post by August Lilleaas about using the :avet index (which requires an index on the attribute in the datomic schema) that I found somewhat helpful.
My DQL query returns only the FROM object, which is nice if the other object were related, but it isn't.
My Query:
$query = $this->em->createQuery('SELECT c, s FROM MyBundle:Person c, MyBundle:Spot s
JOIN s.geo_data g JOIN g.features f WHERE f.active = true AND
ST_Distance(f.location, c.location) < :distance GROUP BY c, s');
This works perfectly in SQL, giving me all the spots and all the persons within :distance of them. But in DQL, it only returns the person object, and since on the database level they are not related, I have no way to fetch the correct spot.
My database setup is correct, I'm using a PostGIS backend and spots and persons are not related in any way. They just happen to be on the same map and I'm querying for spatial relationships.
According to documentation, it's intended behaviour, from what I read, s is being hydrated, but not returned anywhere at all, good job!
How can I teach DQL to please return me what I told it in SELECT? Where's the "I mean what I say, stop being a smartass" switch?
Doctrine cannot give you both entities if they are not related because if the were related you would get the first entity c where you could get s through the relation.
What you can try is selecting all fields of both entities like
SELECT c.location, ..., s.geo_data, ...
This will give you an array for each column that contains all fields from both entities.
Maybe you can use result set mapping to get the entities if desired.
If you want to stuck with Doctrine, you HAVE TO define a OneToMany relation between places and people. In this way, you could set up the PeopleRepository and set up a method like getPeopleByLocationAndMaxDistance(Location $location, $distance)
SELECT p
FROM People AS p
LEFT JOIN Places AS pl
WHERE ST_Distance(p.location, pl.location) < :distance
I need to make a query like this:
SELECT PNPDeviceID FROM Win32_NetworkAdapter WHERE AdapterTypeId = 0
Trouble is, the AdapterTypeId column isn't always present. In this case, I just want everything, like so:
SELECT PNPDeviceID FROM Win32_NetworkAdapter
My WQL/SQL knowledge is extremely limited. Can anybody tell me how to do this in a single query?
EDIT:
A bit more background seems to be required: I am querying Windows for device information using WMI, which uses an SQL-like syntax. So, in my example, I am querying for network adapters that have an AdapterTypeId of 0.
That column is not always present however, meaning that if I enumerate through the returned values then "AdapterTypeId" is not listed.
EDIT 2:
Changed SQL to WQL; apparantly this is more correct.
I am assuming you mean the underlying schema is unreliable.
This is a highly unconventional situation. I suggest that you resolve the issue that is causing the column to not always be present, because to have the schema changing dynamically underneath your application is potentially (almost certainly) disastrous.
Update:
OK, so WQL lets you query objects with a SQL-like syntax but, unlike SQL, the schema can change underneath your feet. This is a classic example of a leaky abstraction, and I now hate WQL without ever having used it :).
Since the available properties are in flux, I am guessing that WQL provides a way to enumerate the properties for a given adapter. Do this, and choose which query to run depending upon the results.
After some Googling, there is an example here, which shows how to enumerate through the available properties. You can use this to determine if AdapterTypeId exists or not.
SELECT PNPDeviceID FROM Win32_NetworkAdapter WHERE AdapterTypeId = {yourDesire} OR AdapterTypeId IS NULL
I assume that you mean that this field is missing from the table.
Do you know before submitting the query if this field exists?
If yes then just create SQL dynamically, otherwise It think you will get syntax error in case of missing field
This is not an SQL question. SQL does not contemplate records with varying schemas in a single table source. Instead (as you mention) this is a different system using an "SQL-like" syntax. You'll have better luck if you recast the question using the actual product that you're trying to query, and information how that product deals with variable record structures is probably discussed in the documentation.