各个输入按号码尾号分区后关联,在输入数据已知分布均匀的情况可以比hash更有效避免倾斜 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525
RAG CheatSheet
langchain+ollama 本地文档常用模型: SBert ollama加载模型在线[src] https://ollama.com/libraryollama run gemma:2b 离线1、创建模型配置文件创建模型配置文件,比如: Modelfile 这个文件名,
Cluster Setup CheatSheet
Cluster-setupBIOS config for Disk SystemDisk Raid 1 DataDisk JBOD OSon demand, ubuntu\fedora\centos\suse\redhat update root passwordsudo pa
Git CheatSheet
检出checkout git checkout 分支名/标签, 该命令会变成detach 只读状态 git checkout -b|-B <new_branch> [<start point>] 基于远程分支名/标签/c
Flink CheatSheet
flinkFlink applications code -> JobGraph -> JobManager -> TaskManagers 环境 ExecutionEnvironment StreamExecutionEnvironment TableEnv
YARN CheatSheet
8088挖矿漏洞发起获取appIDcurl -X POST http://10.33.21.190:8088/ws/v1/cluster/apps/new-application 新建任务信息文件1.json反弹shell{‘application-id’: ‘applicati
Apache Arrow CheatSheet
相关概念包括ValueVector、Field、Schema、VectorSchemaRoot以及Table 1234567891011<dependency> <groupId>org.apache.arrow</groupId>
ClickHouse CheatSheet
启动默认绑定端口9000 与hdfs冲突,修改tcp_port默认配置文件/etc/clickhouse-server/config.xml自定义配置文件目录 /etc/clickhouse-server/config.d/ clickhouse startclickhouse
Oceanbase
JDBC连接Maven Repository: com.oceanbase » oceanbase-client (mvnrepository.com) 引入oceanbase-client-1.1.10.jar到spark的jars目录,使用Beeline连接!connnect
sqlite3
usageyum install sqlite 123456789101112sqlite3 test.db.header oncreate table stu(id int, name char, sex char , score int);insert into stu va