SQLAlchemy的ORM是一个映射函数(Mapper),将Python中定义的类与数据库中的表建立关联,以及类的实例(instance)和表的行(row)建立关联。
查看一个类所对应的数据库表,使用__tablename__属性,例如 User.__tablename__
1. 查询数据 (query)
1.1 查询一个trace中flow个数(to count flows of specific trace)
session.query(Flow).filter(Flow.trace_id == 1).count()
1.2. 查询一个trace中不同srcIP的个数 (to count distinct srcIP)
from sqlalchemy import distinct
from config import *
session = DBSession()
session.query(Flow.srcIP).filter(Flow.trace_id == 1).distinct().count()
1.3 查询一个trace中不同的dstIP和dstPort对的个数(to count distinct dstIP and dstPort)
session.query(Flow.dstIP, Flow.dstPort).filter(Flow.trace_id == 1).distinct().count()
1.4 查询指定列的数据,返回一个KeyedTuple数据类型的列表( get a tuple list of specified columns )
n = session.query(Flow.dstIP, Flow.dstPort).filter(Flow.trace_id == 1).all()
# The type of n is list.
# The type of n[0] is sqlalchemy.util._collections.KeyedTuple
1.5 查询指定列中的所有不同值( get a distinct tuple list of specified columns)
n = session.query(Flow.dstIP, Flow.dstPort).filter(Flow.trace_id == 1).distinct().all()
1.6 获得一列数据的平均值(get average value of a column)
# sql language: select avg(txPkt) from Flow
from sqlalchemy.sql import func
q = session.query(func.avg(Flow.txPkt)).filter(Flow.trace_id == 1)
print q[0][0]
# The type of q is sqlalchemy.orm.query.Query
# The type of q[0] is sqlalchemy.util._collections.KeyedTuple
# The type of q[0][0] is decimal.Decimal
1.7 多列数据平均值的计算(compute average values of columns)
q = session.query((func.avg(Flow.txPkt)+func.avg(Flow.rxPkt))/2).filter(Flow.trace_id == 1)
1.8 对查询到的数据排序(order by )
from sqlalchemy import desc
q = session.query(Flow.timestamp).filter(trace_id == 1).order_by(desc(Flow.timestamp))
1.9 分组查询
q = session.query(Flow.dstIP, Flow.dstPort, func.count(Flow.id)).filter(Flow.trace_id == tid).group_by(Flow.dstIP, Flow.dstPort).all()
2 查询中,常用的过滤操作
等于(equals), 例如 query.filter(name == 'Jack')
不等于(not equals), 例如 query.filter(name != 'Jack')
在列表中(in), 例如 query.filter(name.in_(['Micheal', 'Bob', 'Jack']))
不在列表中(not in), 例如query.filter(~name.in_(['Micheal', 'Bob', 'Jack']))
空值(null), 例如 query.filter(name == None)
不是空值(not null), 例如 query.filter(name != None)
与(and), 例如 query.filter(and_(name == 'Andy', fullname == 'Andy Liu' ))
and_可以省略, 例如 query.filter(name=='Andy', fullname==‘Andy Liu')
或(or), 例如 query.filter(or_(name == 'Andy', name == 'Micheal'))
2. 表的数据操作(table data operation)
2.1 添加删除一个column ( add a new column to a table)
from db import engine
from sqlalchemy import DDL
add_column = DDL('alter table Flow add column cluster_id integer after trace_id')
drop_column = DDL('alter table Flow drop column microsecond')
engine.execute(add_column)
engine.execute(drop_column)
2.2 修改一个数据(update a value)
session.query(Flow).filter(Flow.dstIP == dstIP, Flow.dstPort == dstPort, Flow.trace_id == 1).update({'cluster_id' : 0})
2.3 插入一行数据(insert a row)
session = DBSession()
cluster = Clusters(trace_id = tid, cluster_id = cid,
dstIP = dIP, dstPort = dPort,
avgPkt = aPkt, avgByte = aByte,
size = count)
session.add(cluster)
session.commit() # commit or flush
session.close()
2.4 删除一行数据(delete a row )
session = DBSession()
session.query(Clusters).filter(Clusters.trace_id = 2).delete()
session.commit() # commit or flush
session.close()
补充:
外键 ForeignKey只能引用外表的指定列中已经存在的值。