Is it possible to remove records from crossfilter cube after adding. I need to update the data of crossfilter cube without creating a new cube, from the data that is returned from my API server.
Crossfilter version 1.3 supports removing all records that match the current filters.
Related
The current Kusto data retention policy is mainly based on ingestion time. I am wondering if there is a way to define a data retention policy that is based on some other condition, or any way to mimic the behavior of a conditional data retention policy would do.
For example, I want to remove an item in the database only if there is a newer version of the same (defined by an ID column, say) item got ingested; Otherwise, I'd want the item to be kept in the database regardless its ingestion time. How to achieve that?
I want to remove an item in the database only if there is a newer version of the same (defined by an ID column, say) item got ingested
You could consider creating a materialized view that uses summarize arg_max(version_column, *) by id_column - older records won't be dropped, but if you query the view instead of the table that has the raw data, they will not be visible in your queries.
According to this documentation Here
If I use projection queries with all properties in entity.It will cost me 1 entity read for query and small operation for results.
Is it better than I get all key with key-only queries then get entity data with get(key)? It will cost me 1 entity read for query and N times entity read for entity data.
Thank you.
Note that while projecting will result in a single read op, you need to have all the fields you want to project to be present in the index. An additional index has storage cost and also potentially increase the write latency. So, if you are projecting just a couple of fields of small size, then you can create such a composite index and do a projection and it will only cost one operation.
If composite index with projection is not an option, then you can still try to resolve as much of your where clause as possible via the index and at that point a single query that fetches all the entities will only cost N and (not 1+N where you first get keys and then the entities).
I'm developing a Location Tracking App with Android Studio and need to store coordinates in a normalized manner (One Variable in One Field). Can I modify the location data created by GeoFire in Firebase? I would like to have a single field called Latitude and another one called Longitude. In the picture above you can see that the coordinates are both stored below one field called ' l ', I would like to modify this, or copy them as separate childs of the current userID in the node Location. Would greatly appreciate any help, thanks!
There's nothing built into Geofire for you control the property names. But since Geofire is open-source, you can modify it to use the property names you want.
Just be aware that Geofire uses these shorter names to limit the bandwidth it uses, so your users see a bandwidth increase.
I'd like to query my data with a query that will return me everything on a radius (geofire can find that alright) but also within a datetime window.
At the moment I'm storing datetimes as priorities, so it's quite easy to query the array asking for data between 2 priority numbers (corresponding to start/end datetimes).
It's also quite easy to put my data in a GeoFire array and then query it to get the radius.
Can I combine those 2 though? In an easy not too hacky way?
Cheers
Not with a single query. You must either filter client side, or do your query in more than one phase.
This is because:
Each node is limited to a single priority value.
Per this blog post, GeoFire uses the priority to store a geohash, which it uses for the lookups.
The easiest way to deal with this is to do additional filtering client side. If the bandwidth impact starts causing issues, partition your data (e.g. group events by month) and do it in multiple phases.
I have an existing collection of PERSON's records already loaded to my Solr Server. Each record has a field for a ZIPCODE. I know that Solr now supports spatial search with the geodist() function, the problem is that I do not have the lat and long for each person.
Can I add another collection to Solr mapping the ZipCodes to LATs and LONGs then JOIN them like you would with SQL? Is this even possible with Solr?
AFAIK, there isn't a way to translate Zipcode to Lat/Long in Solr.
Most Geospatial queries (not restricted to Solr) use Latitude and Longitude to perform a radius search. So this isn't a Solr specific problem.
I would recommend enriching the data that is imported into Solr using a GeoCoding API.
You can update your existing index to have these fields populated for every document, but if possible I would prefer recreating the Solr Index with this data.
I wouldn't include another collection for this purpose, JOIN isn't supported in Solr, as it isn't a relational data store.
Edit:
Solr does support JOIN, but in your case I would still not go for it, unless I have to.