Is HIVE database features/dialect enough for Tableau, or is Cloudera Impala better suited for this?

Question

big data

We plan to raise the Hadoop cluster, put project activity data there, followed by realtime (or so) visualization of the main product indicators.
Is the HIVE database features/dialect enough for analytics using the Tableau tool, or is Cloudera Impala better suited for this?
Impala communicates a full-fledged SQL dialect on large amounts of data, but it is very demanding on RAM. Is it worth investing in extra equipment? means or it's not worth it (Not fast enough fetches, very unstable, etc.) and it's better to migrate data from HDFS to Hive, HBase PostgreSQL, etc. in a shrinking state?

Answer 1

So far I got this answer from other sources:
Impala is several times faster, but all the goodies of working with it are limited in the free version, so choosing HIVE still makes sense.
Regarding realtime, neither HBASE nor Impala will give realtime speed by design (they have map reduce under the hood). But realtime speeds can be guaranteed by HBASE and Spark Streaming, or if already prepared data marts are added and distributed from Postgress.