Trying out the osm2pgsql tool, which utilizes copies. Thus far I have found it spends most of its time in the optimizer trying to copy over ~10M rows (~4 of these in parallel). Here is a CPU profile over 10s during the COPY period:

Repro steps:

• Checkout osm2pgsql-dev/osm2pgsql@master...otan-cockroach:crdb_branch, build using cd build && cmake .. && make && cd ..
• Run cockroach single node built with sql: allow escape of any character for COPY #52068 included.
• Download the dataset: https://download.geofabrik.de/europe/iceland-latest.osm.pbf
• Run the tool ./build/osm2pgsql -U root -d defaultdb -H localhost -P 26257 --number-processes 1 -C 20480 ~/iceland-latest.osm.pbf
• Profile during this step:

Processing: Node(6965k 6965.0k/s) Way(356k 14.24k/s) Relation(5290 661.2/s)

PostgreSQL is able to do this in ~7s, we take ~49s. (maybe fodder for https://accidentallyquadratic.tumblr.com/)

Jira issue: CRDB-3984

Epic CRDB-14477
Epic CRDB-14916