Skip to content

Commit 073d1e1

Browse files
committed
Add pg_clickhouse
It dupes the clickhouse benchmark config to load data (except that column names are lowercase) and the Postgres config to run the queries. All queries push down to ClickHouse without error. Use the `TOTAL_PARTITIONS` environment variable to limit the number of partitions to load into ClickHouse, and use `EXPLAIN=1` to emit an explain plan for each query.
1 parent f4b9264 commit 073d1e1

11 files changed

Lines changed: 700 additions & 0 deletions

pg_clickhouse/README.md

Lines changed: 34 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,34 @@
1+
pg_clickhouse ClickBench
2+
========================
3+
4+
This ClickBench configuration benchmarks [pg_clickhouse] running inside
5+
PostgreSQL and pushing queries down to a ClickHouse database. Files:
6+
7+
* `benchmark.sh`: Runs the benchmark
8+
* `clickhouse.sh`: Installs ClickHouse and loads it with data; based on the
9+
[ClickHouse benchmark script](../clickhouse/benchmark.sh)
10+
* `create-postgres.sql`: Loads pg_clickhouse into Postgres and creates the
11+
foreign table
12+
* `create-tuned-memory.sql`, `create-tuned.sql`, and `create.sql`: Scripts
13+
that create the ClickHouse table, copied from the [ClickHouse
14+
Config](../clickhouse/)
15+
* `postgres.sh`: Installs, configures, and starts PostgreSQL; based on the
16+
[PostgreSQL benchmark script](../postgresql/benchmark.sh)
17+
* `queries.sql`: The benchmark queries, identical to the [PostgreSQL
18+
queries](../postgresql/queries.sql)
19+
* `README.md`: This file
20+
* `run.sh`: Runs the benchmark; identical too the [PostgreSQL run
21+
script](../postgresql/run.sh)
22+
* `template.json`: Describes this benchmark configuration
23+
24+
## Environment Variables
25+
26+
### TOTAL_PARTITIONS
27+
28+
Set `$TOTAL_PARTITIONS` to the number of partitions to load into ClickHouse.
29+
Defaults to 100.
30+
31+
### EXPLAIN
32+
33+
Set `$EXPLAIN` to a true value to run each query with
34+
`EXPLAIN (ANALYZE VERBOSE)`.

pg_clickhouse/benchmark.sh

Lines changed: 18 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,18 @@
1+
#!/bin/bash
2+
3+
# apt-get update -y
4+
# env DEBIAN_FRONTEND=noninteractive apt-get install -y wget curl sudo
5+
# env TOTAL_PARTITIONS=1 EXPLAIN=1 ./benchmark.sh
6+
7+
# Install and start ClickHouse and Postgres
8+
./clickhouse.sh "$@"
9+
./postgres.sh
10+
11+
# Run the queries
12+
./run.sh 2>&1 | tee log.txt
13+
14+
echo -n "Data size: "
15+
clickhouse-client --query "SELECT total_bytes FROM system.tables WHERE name = 'hits' AND database = 'default'"
16+
17+
cat log.txt | grep -oP '^Time: \d+\.\d+ ms|psql: error' | sed -r -e 's/Time: ([0-9]+\.[0-9]+) ms/\1/; s/^.*psql: error.*$/null/' |
18+
awk '{ if (i % 3 == 0) { printf "[" }; if ($1 == "null") { printf $1 } else { printf $1 / 1000 }; if (i % 3 != 2) { printf "," } else { print "]," }; ++i; }'

pg_clickhouse/clickhouse.sh

Lines changed: 51 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,51 @@
1+
#!/bin/bash
2+
3+
# Install
4+
5+
if [ ! -x /usr/bin/clickhouse ]
6+
then
7+
cd /tmp || exit
8+
curl https://clickhouse.com/ | sh
9+
sudo ./clickhouse install --noninteractive
10+
rm clickhouse
11+
cd - || exit
12+
fi
13+
14+
# Optional: if you want to use higher compression:
15+
if (( 0 )); then
16+
echo "
17+
compression:
18+
case:
19+
method: zstd
20+
" | sudo tee /etc/clickhouse-server/config.d/compression.yaml
21+
fi;
22+
23+
sudo clickhouse start
24+
25+
for _ in {1..300}
26+
do
27+
clickhouse-client --query "SELECT 1" && break
28+
sleep 1
29+
done
30+
31+
# Determine which set of files to use depending on the type of run
32+
if [ "$1" != "" ] && [ "$1" != "tuned" ] && [ "$1" != "tuned-memory" ]; then
33+
echo "Error: command line argument must be one of {'', 'tuned', 'tuned-memory'}"
34+
exit 1
35+
elif [ ! -z "$1" ]; then
36+
SUFFIX="-$1"
37+
fi
38+
39+
# Load the data
40+
41+
clickhouse-client < create"$SUFFIX".sql
42+
43+
TOTAL_PARTITIONS=${TOTAL_PARTITIONS:-100}
44+
45+
seq 0 "$((TOTAL_PARTITIONS-1))" | xargs -P100 -I{} bash -c 'wget --continue --progress=dot:giga https://datasets.clickhouse.com/hits_compatible/athena_partitioned/hits_{}.parquet'
46+
mkdir -p /var/lib/clickhouse/user_files
47+
sudo mv hits_*.parquet /var/lib/clickhouse/user_files/
48+
sudo chown clickhouse:clickhouse /var/lib/clickhouse/user_files/hits_*.parquet
49+
50+
echo -n "Load time: "
51+
clickhouse-client --time --query "INSERT INTO hits SELECT * FROM file('hits_*.parquet')" --max-insert-threads $(( $(nproc) / 4 ))

pg_clickhouse/create-postgres.sql

Lines changed: 121 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,121 @@
1+
BEGIN;
2+
3+
CREATE EXTENSION IF NOT EXISTS pg_clickhouse;
4+
CREATE SERVER ch_srv FOREIGN DATA WRAPPER clickhouse_fdw
5+
OPTIONS(driver 'binary', host 'localhost', dbname 'default');
6+
CREATE USER MAPPING FOR CURRENT_USER SERVER ch_srv
7+
OPTIONS (user 'default');
8+
9+
CREATE FOREIGN TABLE hits (
10+
WatchID BIGINT NOT NULL,
11+
JavaEnable SMALLINT NOT NULL,
12+
Title TEXT NOT NULL,
13+
GoodEvent SMALLINT NOT NULL,
14+
EventTime TIMESTAMP NOT NULL,
15+
EventDate Date NOT NULL,
16+
CounterID INTEGER NOT NULL,
17+
ClientIP INTEGER NOT NULL,
18+
RegionID INTEGER NOT NULL,
19+
UserID BIGINT NOT NULL,
20+
CounterClass SMALLINT NOT NULL,
21+
OS SMALLINT NOT NULL,
22+
UserAgent SMALLINT NOT NULL,
23+
URL TEXT NOT NULL,
24+
Referer TEXT NOT NULL,
25+
IsRefresh SMALLINT NOT NULL,
26+
RefererCategoryID SMALLINT NOT NULL,
27+
RefererRegionID INTEGER NOT NULL,
28+
URLCategoryID SMALLINT NOT NULL,
29+
URLRegionID INTEGER NOT NULL,
30+
ResolutionWidth SMALLINT NOT NULL,
31+
ResolutionHeight SMALLINT NOT NULL,
32+
ResolutionDepth SMALLINT NOT NULL,
33+
FlashMajor SMALLINT NOT NULL,
34+
FlashMinor SMALLINT NOT NULL,
35+
FlashMinor2 TEXT NOT NULL,
36+
NetMajor SMALLINT NOT NULL,
37+
NetMinor SMALLINT NOT NULL,
38+
UserAgentMajor SMALLINT NOT NULL,
39+
UserAgentMinor VARCHAR(255) NOT NULL,
40+
CookieEnable SMALLINT NOT NULL,
41+
JavascriptEnable SMALLINT NOT NULL,
42+
IsMobile SMALLINT NOT NULL,
43+
MobilePhone SMALLINT NOT NULL,
44+
MobilePhoneModel TEXT NOT NULL,
45+
Params TEXT NOT NULL,
46+
IPNetworkID INTEGER NOT NULL,
47+
TraficSourceID SMALLINT NOT NULL,
48+
SearchEngineID SMALLINT NOT NULL,
49+
SearchPhrase TEXT NOT NULL,
50+
AdvEngineID SMALLINT NOT NULL,
51+
IsArtifical SMALLINT NOT NULL,
52+
WindowClientWidth SMALLINT NOT NULL,
53+
WindowClientHeight SMALLINT NOT NULL,
54+
ClientTimeZone SMALLINT NOT NULL,
55+
ClientEventTime TIMESTAMP NOT NULL,
56+
SilverlightVersion1 SMALLINT NOT NULL,
57+
SilverlightVersion2 SMALLINT NOT NULL,
58+
SilverlightVersion3 INTEGER NOT NULL,
59+
SilverlightVersion4 SMALLINT NOT NULL,
60+
PageCharset TEXT NOT NULL,
61+
CodeVersion INTEGER NOT NULL,
62+
IsLink SMALLINT NOT NULL,
63+
IsDownload SMALLINT NOT NULL,
64+
IsNotBounce SMALLINT NOT NULL,
65+
FUniqID BIGINT NOT NULL,
66+
OriginalURL TEXT NOT NULL,
67+
HID INTEGER NOT NULL,
68+
IsOldCounter SMALLINT NOT NULL,
69+
IsEvent SMALLINT NOT NULL,
70+
IsParameter SMALLINT NOT NULL,
71+
DontCountHits SMALLINT NOT NULL,
72+
WithHash SMALLINT NOT NULL,
73+
HitColor CHAR NOT NULL,
74+
LocalEventTime TIMESTAMP NOT NULL,
75+
Age SMALLINT NOT NULL,
76+
Sex SMALLINT NOT NULL,
77+
Income SMALLINT NOT NULL,
78+
Interests SMALLINT NOT NULL,
79+
Robotness SMALLINT NOT NULL,
80+
RemoteIP INTEGER NOT NULL,
81+
WindowName INTEGER NOT NULL,
82+
OpenerName INTEGER NOT NULL,
83+
HistoryLength SMALLINT NOT NULL,
84+
BrowserLanguage TEXT NOT NULL,
85+
BrowserCountry TEXT NOT NULL,
86+
SocialNetwork TEXT NOT NULL,
87+
SocialAction TEXT NOT NULL,
88+
HTTPError SMALLINT NOT NULL,
89+
SendTiming INTEGER NOT NULL,
90+
DNSTiming INTEGER NOT NULL,
91+
ConnectTiming INTEGER NOT NULL,
92+
ResponseStartTiming INTEGER NOT NULL,
93+
ResponseEndTiming INTEGER NOT NULL,
94+
FetchTiming INTEGER NOT NULL,
95+
SocialSourceNetworkID SMALLINT NOT NULL,
96+
SocialSourcePage TEXT NOT NULL,
97+
ParamPrice BIGINT NOT NULL,
98+
ParamOrderID TEXT NOT NULL,
99+
ParamCurrency TEXT NOT NULL,
100+
ParamCurrencyID SMALLINT NOT NULL,
101+
OpenstatServiceName TEXT NOT NULL,
102+
OpenstatCampaignID TEXT NOT NULL,
103+
OpenstatAdID TEXT NOT NULL,
104+
OpenstatSourceID TEXT NOT NULL,
105+
UTMSource TEXT NOT NULL,
106+
UTMMedium TEXT NOT NULL,
107+
UTMCampaign TEXT NOT NULL,
108+
UTMContent TEXT NOT NULL,
109+
UTMTerm TEXT NOT NULL,
110+
FromTag TEXT NOT NULL,
111+
HasGCLID SMALLINT NOT NULL,
112+
RefererHash BIGINT NOT NULL,
113+
URLHash BIGINT NOT NULL,
114+
CLID INTEGER NOT NULL
115+
) SERVER ch_srv OPTIONS(
116+
database 'default',
117+
table_name 'hits',
118+
engine 'MergeTree'
119+
);
120+
121+
COMMIT;
Lines changed: 109 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,109 @@
1+
CREATE OR REPLACE TABLE hits
2+
(
3+
watchid BIGINT NOT NULL,
4+
javaenable SMALLINT NOT NULL,
5+
title TEXT NOT NULL,
6+
goodevent SMALLINT NOT NULL,
7+
eventtime TIMESTAMP NOT NULL,
8+
eventdate Date NOT NULL,
9+
counterid INTEGER NOT NULL,
10+
clientip INTEGER NOT NULL,
11+
regionid INTEGER NOT NULL,
12+
userid BIGINT NOT NULL,
13+
counterclass SMALLINT NOT NULL,
14+
os SMALLINT NOT NULL,
15+
useragent SMALLINT NOT NULL,
16+
url TEXT NOT NULL,
17+
referer TEXT NOT NULL,
18+
isrefresh SMALLINT NOT NULL,
19+
referercategoryid SMALLINT NOT NULL,
20+
refererregionid INTEGER NOT NULL,
21+
urlcategoryid SMALLINT NOT NULL,
22+
urlregionid INTEGER NOT NULL,
23+
resolutionwidth SMALLINT NOT NULL,
24+
resolutionheight SMALLINT NOT NULL,
25+
resolutiondepth SMALLINT NOT NULL,
26+
flashmajor SMALLINT NOT NULL,
27+
flashminor SMALLINT NOT NULL,
28+
flashminor2 TEXT NOT NULL,
29+
netmajor SMALLINT NOT NULL,
30+
netminor SMALLINT NOT NULL,
31+
useragentmajor SMALLINT NOT NULL,
32+
useragentminor VARCHAR(255) NOT NULL,
33+
cookieenable SMALLINT NOT NULL,
34+
javascriptenable SMALLINT NOT NULL,
35+
ismobile SMALLINT NOT NULL,
36+
mobilephone SMALLINT NOT NULL,
37+
mobilephonemodel TEXT NOT NULL,
38+
params TEXT NOT NULL,
39+
ipnetworkid INTEGER NOT NULL,
40+
traficsourceid SMALLINT NOT NULL,
41+
searchengineid SMALLINT NOT NULL,
42+
searchphrase TEXT NOT NULL,
43+
advengineid SMALLINT NOT NULL,
44+
isartifical SMALLINT NOT NULL,
45+
windowclientwidth SMALLINT NOT NULL,
46+
windowclientheight SMALLINT NOT NULL,
47+
clienttimezone SMALLINT NOT NULL,
48+
clienteventtime TIMESTAMP NOT NULL,
49+
silverlightversion1 SMALLINT NOT NULL,
50+
silverlightversion2 SMALLINT NOT NULL,
51+
silverlightversion3 INTEGER NOT NULL,
52+
silverlightversion4 SMALLINT NOT NULL,
53+
pagecharset TEXT NOT NULL,
54+
codeversion INTEGER NOT NULL,
55+
islink SMALLINT NOT NULL,
56+
isdownload SMALLINT NOT NULL,
57+
isnotbounce SMALLINT NOT NULL,
58+
funiqid BIGINT NOT NULL,
59+
originalurl TEXT NOT NULL,
60+
hid INTEGER NOT NULL,
61+
isoldcounter SMALLINT NOT NULL,
62+
isevent SMALLINT NOT NULL,
63+
isparameter SMALLINT NOT NULL,
64+
dontcounthits SMALLINT NOT NULL,
65+
withhash SMALLINT NOT NULL,
66+
hitcolor CHAR NOT NULL,
67+
localeventtime TIMESTAMP NOT NULL,
68+
age SMALLINT NOT NULL,
69+
sex SMALLINT NOT NULL,
70+
income SMALLINT NOT NULL,
71+
interests SMALLINT NOT NULL,
72+
robotness SMALLINT NOT NULL,
73+
remoteip INTEGER NOT NULL,
74+
windowname INTEGER NOT NULL,
75+
openername INTEGER NOT NULL,
76+
historylength SMALLINT NOT NULL,
77+
browserlanguage TEXT NOT NULL,
78+
browsercountry TEXT NOT NULL,
79+
socialnetwork TEXT NOT NULL,
80+
socialaction TEXT NOT NULL,
81+
httperror SMALLINT NOT NULL,
82+
sendtiming INTEGER NOT NULL,
83+
dnstiming INTEGER NOT NULL,
84+
connecttiming INTEGER NOT NULL,
85+
responsestarttiming INTEGER NOT NULL,
86+
responseendtiming INTEGER NOT NULL,
87+
fetchtiming INTEGER NOT NULL,
88+
socialsourcenetworkid SMALLINT NOT NULL,
89+
socialsourcepage TEXT NOT NULL,
90+
paramprice BIGINT NOT NULL,
91+
paramorderid TEXT NOT NULL,
92+
paramcurrency TEXT NOT NULL,
93+
paramcurrencyid SMALLINT NOT NULL,
94+
openstatservicename TEXT NOT NULL,
95+
openstatcampaignid TEXT NOT NULL,
96+
openstatadid TEXT NOT NULL,
97+
openstatsourceid TEXT NOT NULL,
98+
utmsource TEXT NOT NULL,
99+
utmmedium TEXT NOT NULL,
100+
utmcampaign TEXT NOT NULL,
101+
utmcontent TEXT NOT NULL,
102+
utmterm TEXT NOT NULL,
103+
fromtag TEXT NOT NULL,
104+
hasgclid SMALLINT NOT NULL,
105+
refererhash BIGINT NOT NULL,
106+
urlhash BIGINT NOT NULL,
107+
clid INTEGER NOT NULL,
108+
)
109+
ENGINE = Memory;

0 commit comments

Comments
 (0)