distribute_agg_avg.py

from util.log import *
from util.cases import *
from util.sql import *
import numpy as np
import random ,os ,sys
import platform


class TDTestCase:

    updatecfgDict = {"maxTablesPerVnode":2 ,"minTablesPerVnode":2,"tableIncStepPerVnode":2 }
    def init(self, conn, logSql):
        tdLog.debug("start to execute %s" % __file__)
        tdSql.init(conn.cursor())
        self.vnode_disbutes = None
        self.ts = 1537146000000


    def check_avg_functions(self, tbname , col_name):

        avg_sql = f"select avg({col_name}) from {tbname};"

        same_sql = f"select {col_name} from {tbname} where {col_name} is not null "

        tdSql.query(same_sql)
        pre_data = np.array(tdSql.queryResult)[np.array(tdSql.queryResult) != None]
        if (platform.system().lower() == 'windows' and pre_data.dtype == 'int32'):
            pre_data = np.array(pre_data, dtype = 'int64')
        pre_avg = np.sum(pre_data)/len(pre_data)

        tdSql.query(avg_sql)
        tdSql.checkData(0,0,pre_avg)

    def prepare_datas_of_distribute(self, dbname="testdb"):

        # prepate datas for  20 tables distributed at different vgroups
        tdSql.execute(f"create database if not exists {dbname} keep 3650 duration 1000 vgroups 5")
        tdSql.execute(f" use {dbname} ")
        tdSql.execute(
            f'''create table {dbname}.stb1
            (ts timestamp, c1 int, c2 bigint, c3 smallint, c4 tinyint, c5 float, c6 double, c7 bool, c8 binary(16),c9 nchar(32), c10 timestamp)
            tags (t0 timestamp, t1 int, t2 bigint, t3 smallint, t4 tinyint, t5 float, t6 double, t7 bool, t8 binary(16),t9 nchar(32))
            '''
        )

        for i in range(20):
            tdSql.execute(f'create table {dbname}.ct{i+1} using {dbname}.stb1 tags ( now(), {1*i}, {11111*i}, {111*i}, {1*i}, {1.11*i}, {11.11*i}, {i%2}, "binary{i}", "nchar{i}" )')

        for i in range(9):
            tdSql.execute(
                f"insert into {dbname}.ct1 values ( now()-{i*10}s, {1*i}, {11111*i}, {111*i}, {11*i}, {1.11*i}, {11.11*i}, {i%2}, 'binary{i}', 'nchar{i}', now()+{1*i}a )"
            )
            tdSql.execute(
                f"insert into {dbname}.ct4 values ( now()-{i*90}d, {1*i}, {11111*i}, {111*i}, {11*i}, {1.11*i}, {11.11*i}, {i%2}, 'binary{i}', 'nchar{i}', now()+{1*i}a )"
            )

        for i in range(1,21):
            if i ==1 or i == 4:
                continue
            else:
                tbname = f"{dbname}.ct{i}"
                for j in range(9):
                    tdSql.execute(
                f"insert into {tbname} values ( now()-{(i+j)*10}s, {1*(j+i)}, {11111*(j+i)}, {111*(j+i)}, {11*(j)}, {1.11*(j+i)}, {11.11*(j+i)}, {(j+i)%2}, 'binary{j}', 'nchar{j}', now()+{1*j}a )"
            )
        tdSql.execute(f"insert into {dbname}.ct1 values (now()-45s, 0, 0, 0, 0, 0, 0, 0, 'binary0', 'nchar0', now()+8a )")
        tdSql.execute(f"insert into {dbname}.ct1 values (now()+10s, 9, -99999, -999, -99, -9.99, -99.99, 1, 'binary9', 'nchar9', now()+9a )")
        tdSql.execute(f"insert into {dbname}.ct1 values (now()+15s, 9, -99999, -999, -99, -9.99, NULL, 1, 'binary9', 'nchar9', now()+9a )")
        tdSql.execute(f"insert into {dbname}.ct1 values (now()+20s, 9, -99999, -999, NULL, -9.99, -99.99, 1, 'binary9', 'nchar9', now()+9a )")

        tdSql.execute(f"insert into {dbname}.ct4 values (now()-810d, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL ) ")
        tdSql.execute(f"insert into {dbname}.ct4 values (now()-400d, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL ) ")
        tdSql.execute(f"insert into {dbname}.ct4 values (now()+90d, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL  ) ")

        tdLog.info(" prepare data for distributed_aggregate done! ")

    def check_distribute_datas(self, dbname="testdb"):
        # get vgroup_ids of all
        tdSql.query(f"show {dbname}.vgroups ")
        vgroups = tdSql.queryResult

        vnode_tables={}

        for vgroup_id in vgroups:
            vnode_tables[vgroup_id[0]]=[]


        # check sub_table of per vnode ,make sure sub_table has been distributed
        tdSql.query(f"show {dbname}.tables like 'ct%'")
        table_names = tdSql.queryResult
        tablenames = []
        for table_name in table_names:
            vnode_tables[table_name[6]].append(table_name[0])
        self.vnode_disbutes = vnode_tables

        count = 0
        for k ,v in vnode_tables.items():
            if len(v)>=2:
                count+=1
        if count < 2:
            tdLog.exit(" the datas of all not satisfy sub_table has been distributed ")

    def check_avg_distribute_diff_vnode(self,col_name, dbname="testdb"):

        vgroup_ids = []
        for k ,v in self.vnode_disbutes.items():
            if len(v)>=2:
                vgroup_ids.append(k)

        distribute_tbnames = []

        for vgroup_id in vgroup_ids:
            vnode_tables = self.vnode_disbutes[vgroup_id]
            distribute_tbnames.append(random.sample(vnode_tables,1)[0])
        tbname_ins = ""
        for tbname in distribute_tbnames:
            tbname_ins += "'%s' ,"%tbname

        tbname_filters = tbname_ins[:-1]

        avg_sql = f"select avg({col_name}) from {dbname}.stb1 where tbname in ({tbname_filters});"

        same_sql = f"select {col_name}  from {dbname}.stb1 where tbname in ({tbname_filters}) and {col_name} is not null "

        tdSql.query(same_sql)
        pre_data = np.array(tdSql.queryResult)[np.array(tdSql.queryResult) != None]
        if (platform.system().lower() == 'windows' and pre_data.dtype == 'int32'):
            pre_data = np.array(pre_data, dtype = 'int64')
        pre_avg = np.sum(pre_data)/len(pre_data)

        tdSql.query(avg_sql)
        tdSql.checkData(0,0,pre_avg)

    def check_avg_status(self, dbname="testdb"):
        # check max function work status

        tdSql.query(f"show {dbname}.tables like 'ct%'")
        table_names = tdSql.queryResult
        tablenames = []
        for table_name in table_names:
            tablenames.append(f"{dbname}.{table_name[0]}")

        tdSql.query(f"desc {dbname}.stb1")
        col_names = tdSql.queryResult

        colnames = []
        for col_name in col_names:
            if col_name[1] in ["INT" ,"BIGINT" ,"SMALLINT" ,"TINYINT" , "FLOAT" ,"DOUBLE"]:
                colnames.append(col_name[0])

        for tablename in tablenames:
            for colname in colnames:
                self.check_avg_functions(tablename,colname)

        # check max function for different vnode

        for colname in colnames:
            if colname.startswith("c"):
                self.check_avg_distribute_diff_vnode(colname, dbname)
            else:
                # self.check_avg_distribute_diff_vnode(colname, dbname) # bug for tag
                pass


    def distribute_agg_query(self, dbname="testdb"):
        # basic filter
        tdSql.query(f"select avg(c1) from {dbname}.stb1 ")
        tdSql.checkData(0,0,14.086956522)

        tdSql.query(f"select avg(a) from (select avg(c1) a  from {dbname}.stb1 partition by tbname) ")
        tdSql.checkData(0,0,14.292307692)

        tdSql.query(f"select avg(c1) from {dbname}.stb1 where t1=1")
        tdSql.checkData(0,0,6.000000000)

        tdSql.query(f"select avg(c1+c2) from {dbname}.stb1 where c1 =1 ")
        tdSql.checkData(0,0,11112.000000000)

        tdSql.query(f"select avg(c1) from {dbname}.stb1 where tbname=\"ct2\"")
        tdSql.checkData(0,0,6.000000000)

        tdSql.query(f"select avg(c1) from {dbname}.stb1 partition by tbname")
        tdSql.checkRows(20)

        tdSql.query(f"select avg(c1) from {dbname}.stb1 where t1> 4  partition by tbname")
        tdSql.checkRows(15)

        # union all
        tdSql.query(f"select avg(c1) from {dbname}.stb1 union all select avg(c1) from {dbname}.stb1 ")
        tdSql.checkRows(2)
        tdSql.checkData(0,0,14.086956522)

        tdSql.query(f"select avg(a) from (select avg(c1) a from {dbname}.stb1 union all select avg(c1) a  from {dbname}.stb1)")
        tdSql.checkRows(1)
        tdSql.checkData(0,0,14.086956522)

        # join

        tdSql.execute(" create database if not exists db ")
        tdSql.execute(" use db ")
        tdSql.execute(" create stable db.st (ts timestamp , c1 int ,c2 float) tags(t1 int) ")
        tdSql.execute(" create table db.tb1 using db.st tags(1) ")
        tdSql.execute(" create table db.tb2 using db.st tags(2) ")


        for i in range(10):
            ts = i*10 + self.ts
            tdSql.execute(f" insert into db.tb1 values({ts},{i},{i}.0)")
            tdSql.execute(f" insert into db.tb2 values({ts},{i},{i}.0)")

        tdSql.query(f"select avg(tb1.c1), avg(tb2.c2) from db.tb1 tb1, db.tb2 tb2 where tb1.ts=tb2.ts")
        tdSql.checkRows(1)
        tdSql.checkData(0,0,4.500000000)
        tdSql.checkData(0,1,4.500000000)

        # group by
        tdSql.execute(f" use {dbname} ")

        # partition by tbname or partition by tag
        tdSql.query(f"select avg(c1) from {dbname}.stb1 partition by tbname")
        tdSql.checkRows(20)

        # nest query for support max
        tdSql.query(f"select avg(c2+2)+1 from (select avg(c1) c2  from {dbname}.stb1)")
        tdSql.checkData(0,0,17.086956522)
        tdSql.query(f"select avg(c1+2)  as c2 from (select ts ,c1 ,c2  from {dbname}.stb1)")
        tdSql.checkData(0,0,16.086956522)
        tdSql.query(f"select avg(a+2)  as c2 from (select ts ,abs(c1) a ,c2  from {dbname}.stb1)")
        tdSql.checkData(0,0,16.086956522)

        # mixup with other functions
        tdSql.query(f"select max(c1),count(c1),last(c2,c3),sum(c1+c2),avg(c1) from {dbname}.stb1")
        tdSql.checkData(0,0,28)
        tdSql.checkData(0,1,184)
        tdSql.checkData(0,2,-99999)
        tdSql.checkData(0,3,-999)
        tdSql.checkData(0,4,28202310.000000000)
        tdSql.checkData(0,5,14.086956522)

    def run(self):

        self.prepare_datas_of_distribute()
        self.check_distribute_datas()
        self.check_avg_status()
        self.distribute_agg_query()


    def stop(self):
        tdSql.close()
        tdLog.success("%s successfully executed" % __file__)

tdCases.addWindows(__file__, TDTestCase())
tdCases.addLinux(__file__, TDTestCase())