feat(script): generate tpch data set

.gitignore

@@ -57,3 +57,6 @@ venv/
 __pycache__/
 
 *.zip
+
+# tpch data set
+tpch/data

.licenserc.yaml

@@ -10,6 +10,7 @@ header:
     - "website"
     - "tests"
     - "tools"
+    - "tpch"
     # Ignore hidden files
     - ".cargo"
     - ".databend"

tpch/run-tpch-dbgen.sh

@@ -0,0 +1,5 @@
+#!/bin/bash
+
+cd /tpch-dbgen
+./dbgen -vf -s 1
+mv *.tbl /data

tpch/tpch-data.sh

@@ -0,0 +1,14 @@
+#!/bin/bash
+
+# Construct a docker imagine to generate tpch-data
+docker build  -f tpchdata.dockerfile -t databend:latest .
+
+# Generate data into the ./data directory if it does not already exist
+FILE=./data/customer.tbl
+if test -f "$FILE"; then
+    echo "$FILE exists."
+else
+  mkdir data 2>/dev/null
+  docker run -v `pwd`/data:/data --rm databend:latest
+  ls -l data
+fi

tpch/tpchdata.dockerfile

@@ -0,0 +1,14 @@
+FROM ubuntu:22.04
+
+RUN apt-get update && \
+    apt-get install -y git build-essential
+
+# Use https://github.com/databricks/tpch-dbgen to generate data
+RUN git clone https://github.com/databricks/tpch-dbgen.git && cd tpch-dbgen && make
+
+WORKDIR /tpch-dbgen
+ADD run-tpch-dbgen.sh /tpch-dbgen/
+
+VOLUME /data
+
+ENTRYPOINT [ "bash", "./run-tpch-dbgen.sh" ]