Yes, the transformer runs on the server side, so it should be functionally the same. However, a scan is going to copy the key and value to the heap, so it’s likely to use more memory than a SQL query.
> On 8 Mar 2022, at 10:04, Surinder Mehra <redni...@gmail.com> wrote: > > Hi, > Actually SQL is not enabled on these caches. > We can try compute but I was thinking to use scan query with 'transformer' . > I assume transformer runs on server node right so isn't it same as "select > _key from cache" > > On Tue, 8 Mar 2022, 15:08 Stephen Darlington, > <stephen.darling...@gridgain.com <mailto:stephen.darling...@gridgain.com>> > wrote: > You could use SQL. “select _key from table” > > But really, copying all the data over the network is often not the best > strategy. Send a compute job to each node or partition and work on the data > “in place” on the data nodes. > >> On 8 Mar 2022, at 03:56, Surinder Mehra <redni...@gmail.com >> <mailto:redni...@gmail.com>> wrote: >> >> Hi, >> I was looking for a way to fetch all cache keys in an efficient manner from >> ignite cache. Looks like there is no such api yet. >> Would it be correct to use a transformer in Scan query to just fetch the key >> from entry being scanned. This will avoid fetching full cache entry from >> server nodes >> >> Or is there a better way to do this. >> >> https://www.gridgain.com/docs/latest/developers-guide/key-value-api/using-scan-queries >> >> <https://www.gridgain.com/docs/latest/developers-guide/key-value-api/using-scan-queries> >> >