【Springboot】——整合p6spy格式化SQL日志

项目现状介绍

    项目使用Springboot+Mybatis作为基础框架，项目中日志框架采用logback+Slf4J，关于日志中SQL部分日志的打印采用的mybatis本身打印格式，sql语句正常输出，参数的位置使用占位符“？”替代。这样造成问题就是在处理项目中的sql问题，sql格式打印不规范，从sql日志查看，到sql日志运行判断都会花费一些重复的时间，既然这样那就想办法节省这一部分的时间吧。在Java项目中格式化sql日志，推荐使用辅助jar包，p6spy。

何为p6spy？

    p6spy是针对数据库访问操作的动态监测框架（开源项目）它使得数据库数据可无缝截取和操纵，而不必对现有应用程序的代码作任何修改。P6Spy 分发包包括P6Log，它是一 个可记录任何 Java 应用程序的所有JDBC事务的应用程序。其配置完成使用时，可以进行数据访问性能的监测。
  p6spy是一个可以用来在应用程序中拦截和修改数据操作语句的开源框架。 通过p6spy我们可以对SQL语句进行拦截，相当于一个SQL语句的记录器，这样我们可以用它来作相关的分析，比如性能分析等我们最需要的功能，查看sql语句，不是预编译的带问号的，而是真正的数据库执行的sql，更直观，更简单。
p6spy相关材料的获取途径
项目首页：http://p6spy.github.io/p6spy/
GitHub下载介绍页面：https://github.com/p6spy/p6spy/wiki/Download
GitHub托管地址：https://github.com/p6spy/p6spy/tree/master
帮助文档地址：http://p6spy.readthedocs.io/en/latest/

Springboot+Mybatis整合p6spy

#引入jar包

        p6spy
        p6spy
        3.7.0

###引入配置文件spy.properties
###
# #%L
# P6Spy
# %%
# Copyright (C) 2013 P6Spy
# %%
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#      http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# #L%
###
#################################################################
# P6Spy Options File                                            #
# See documentation for detailed instructions                   #
# http://p6spy.github.io/p6spy/2.0/configandusage.html          #
#################################################################
#################################################################
# MODULES                                                       #
#                                                               #
# Module list adapts the modular functionality of P6Spy.        #
# Only modules listed are active.                               #
# (default is com.p6spy.engine.logging.P6LogFactory and         #
# com.p6spy.engine.spy.P6SpyFactory)                            #
# Please note that the core module (P6SpyFactory) can't be      #
# deactivated.                                                  #
# Unlike the other properties, activation of the changes on     #
# this one requires reload.                                     #
#################################################################
#modulelist=com.p6spy.engine.spy.P6SpyFactory,com.p6spy.engine.logging.P6LogFactory,com.p6spy.engine.outage.P6OutageFactory
################################################################
# CORE (P6SPY) PROPERTIES                                      #
################################################################
# A comma separated list of JDBC drivers to load and register.
# (default is empty)
#
# Note: This is normally only needed when using P6Spy in an
# application server environment with a JNDI data source or when
# using a JDBC driver that does not implement the JDBC 4.0 API
# (specifically automatic registration).
#driverlist=
#####配置数据库驱动，根据自己的数据库选择
driverlist=com.mysql.jdbc.Driver
# for flushing per statement
# (default is false)
#autoflush = false
# sets the date format using Java's SimpleDateFormat routine.
# In case property is not set, miliseconds since 1.1.1970 (unix time) is used (default is empty)
#dateformat=
# prints a stack trace for every statement logged
#stacktrace=false
# if stacktrace=true, specifies the stack trace to print
#stacktraceclass=
# determines if property file should be reloaded
# Please note: reload means forgetting all the previously set
# settings (even those set during runtime - via JMX)
# and starting with the clean table
# (default is false)
#reloadproperties=false
reloadproperties=true
# determines how often should be reloaded in seconds
# (default is 60)
#reloadpropertiesinterval=60
# specifies the appender to use for logging
# Please note: reload means forgetting all the previously set
# settings (even those set during runtime - via JMX)
# and starting with the clean table
# (only the properties read from the configuration file)
# (default is com.p6spy.engine.spy.appender.FileLogger)
###选择sql日志输出的问题 slf4jLogger输出到控制台，fileLogger输出到文件中
appender=com.p6spy.engine.spy.appender.Slf4JLogger
#appender=com.p6spy.engine.spy.appender.StdoutLogger
#appender=com.p6spy.engine.spy.appender.FileLogger
# name of logfile to use, note Windows users should make sure to use forward slashes in their pathname (e:/test/spy.log)
# (used for com.p6spy.engine.spy.appender.FileLogger only)
# (default is spy.log)
###如果选择输出到文件中自定义文件的输出路径
logfile = /data/www/logs/spy.log
# append to the p6spy log file. if this is set to false the
# log file is truncated every time. (file logger only)
# (default is true)
#append=true
# class to use for formatting log messages (default is: com.p6spy.engine.spy.appender.SingleLineFormat)
#logMessageFormat=com.p6spy.engine.spy.appender.MultiLineFormat
#自定义日志格式，在类中定义，因为原生的p6spy日志格式虽然取代了“？”但是全部输出到一行，可读性差，还是建议自定义一下
logMessageFormat=com.xxx.xxxx.common.datasource.P6Spy
#logMessageFormat=com.p6spy.engine.spy.appender.SingleLineFormat
# format that is used for logging of the date/time/... (has to be compatible with java.text.SimpleDateFormat)
# (default is dd-MMM-yy)
#databaseDialectDateFormat=dd-MMM-yy
##输出时间格式
databaseDialectDateFormat=yyyy-MM-dd HH:mm:ss
# whether to expose options via JMX or not
# (default is true)
#jmx=true
# if exposing options via jmx (see option: jmx), what should be the prefix used?
# jmx naming pattern constructed is: com.p6spy(.)?:name=
# please note, if there is already such a name in use it would be unregistered first (the last registered wins)
# (default is none)
#jmxPrefix=
#################################################################
# DataSource replacement                                        #
#                                                               #
# Replace the real DataSource class in your application server  #
# configuration with the name com.p6spy.engine.spy.P6DataSource #
# (that provides also connection pooling and xa support).       #
# then add the JNDI name and class name of the real             #
# DataSource here                                               #
#                                                               #
# Values set in this item cannot be reloaded using the          #
# reloadproperties variable. Once it is loaded, it remains      #
# in memory until the application is restarted.                 #
#                                                               #
#################################################################
#realdatasource=/RealMySqlDS
#realdatasourceclass=com.mysql.jdbc.jdbc2.optional.MysqlDataSource
#################################################################
# DataSource properties                                         #
#                                                               #
# If you are using the DataSource support to intercept calls    #
# to a DataSource that requires properties for proper setup,    #
# define those properties here. Use name value pairs, separate  #
# the name and value with a semicolon, and separate the         #
# pairs with commas.                                            #
#                                                               #
# The example shown here is for mysql                           #
#                                                               #
#################################################################
#realdatasourceproperties=port;3306,serverName;myhost,databaseName;jbossdb,foo;bar
#################################################################
# JNDI DataSource lookup                                        #
#                                                               #
# If you are using the DataSource support outside of an app     #
# server, you will probably need to define the JNDI Context     #
# environment.                                                  #
#                                                               #
# If the P6Spy code will be executing inside an app server then #
# do not use these properties, and the DataSource lookup will   #
# use the naming context defined by the app server.             #
#                                                               #
# The two standard elements of the naming environment are       #
# jndicontextfactory and jndicontextproviderurl. If you need    #
# additional elements, use the jndicontextcustom property.      #
# You can define multiple properties in jndicontextcustom,      #
# in name value pairs. Separate the name and value with a       #
# semicolon, and separate the pairs with commas.                #
#                                                               #
# The example shown here is for a standalone program running on #
# a machine that is also running JBoss, so the JDNI context     #
# is configured for JBoss (3.0.4).                              #
#                                                               #
# (by default all these are empty)                              #
#################################################################
#jndicontextfactory=org.jnp.interfaces.NamingContextFactory
#jndicontextproviderurl=localhost:1099
#jndicontextcustom=java.naming.factory.url.pkgs;org.jboss.nameing:org.jnp.interfaces
#jndicontextfactory=com.ibm.websphere.naming.WsnInitialContextFactory
#jndicontextproviderurl=iiop://localhost:900
################################################################
# P6 LOGGING SPECIFIC PROPERTIES                               #
################################################################
# filter what is logged
# please note this is a precondition for usage of: include/exclude/sqlexpression
# (default is false)
#filter=true
# comma separated list of strings to include
# please note that special characters escaping (used in java) has to be done for the provided regular expression
# (default is empty)
#include=select
# comma separated list of strings to exclude
# (default is empty)
#exclude =
# sql expression to evaluate if using regex
# please note that special characters escaping (used in java) has to be done for the provided regular expression
# (default is empty)
#sqlexpression =
#list of categories to exclude: error, info, batch, debug, statement,
#commit, rollback and result are valid values
# (default is info,debug,result,resultset,batch)
#excludecategories=info,debug,result,resultset,batch
#excludecategories=error,info,debug,result,resultset
#excludecategories=info
# Execution threshold applies to the standard logging of P6Spy.
# While the standard logging logs out every statement
# regardless of its execution time, this feature puts a time
# condition on that logging. Only statements that have taken
# longer than the time specified (in milliseconds) will be
# logged. This way it is possible to see only statements that
# have exceeded some high water mark.
# This time is reloadable.
#
# executionThreshold=integer time (milliseconds)
# (default is 0)
#executionThreshold=
################################################################
# P6 OUTAGE SPECIFIC PROPERTIES                                #
################################################################
# Outage Detection
#
# This feature detects long-running statements that may be indicative of
# a database outage problem. If this feature is turned on, it will log any
# statement that surpasses the configurable time boundary during its execution.
# When this feature is enabled, no other statements are logged except the long
# running statements. The interval property is the boundary time set in seconds.
# For example, if this is set to 2, then any statement requiring at least 2
# seconds will be logged. Note that the same statement will continue to be logged
# for as long as it executes. So if the interval is set to 2, and the query takes
# 11 seconds, it will be logged 5 times (at the 2, 4, 6, 8, 10 second intervals).
#
# outagedetection=true|false
# outagedetectioninterval=integer time (seconds)
#
# (default is false)
#outagedetection=false
# (default is 60)
#outagedetectioninterval=30

自定义日志输出格式代码

import com.p6spy.engine.spy.appender.MessageFormattingStrategy;

import java.text.SimpleDateFormat;
import java.util.Date;

public class P6Spy implements MessageFormattingStrategy {
    private SimpleDateFormat format = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss:SSS");

    @Override
    public String formatMessage(int connectionId, String now, long elapsed, String category, String prepared, String sql) {
        return !"".equals(sql.trim()) ? this.format.format(new Date()) + " | took " + elapsed + "ms | " + category + " | connection " + connectionId + "\n " + sql + ";" : "";
    }
}

修改数据源，因为项目中配置了动态链接数据源，所以在代码中将返回Datasource的地方替换成P6Datasource

 public DataSource createDataSource(DataSourceConfig dataSourceConfig, Map propertyMap) {

        Class dataSourceType;
        try {
            dataSourceType = (Class) Class.forName((String) "org.apache.tomcat.jdbc.pool.DataSource");
            DataSourceBuilder factory = DataSourceBuilder.create().driverClassName("com.mysql.jdbc.Driver").url(dataSourceConfig.getUrl())
                    .username(dataSourceConfig.getUserName()).password(dataSourceConfig.getPassWord()).type(dataSourceType);
            ConversionService conversionService = new DefaultConversionService();
            DataSource dataSource = factory.build();
            RelaxedDataBinder dataBinder = new RelaxedDataBinder(dataSource);
            dataBinder.setConversionService(conversionService);
            dataBinder.setIgnoreNestedProperties(false);//false
            dataBinder.setIgnoreInvalidFields(false);//false
            dataBinder.setIgnoreUnknownFields(true);//true
            PropertyValues dataSourcePropertyValues = new MutablePropertyValues(propertyMap);
            dataBinder.bind(dataSourcePropertyValues);
            return new P6DataSource(dataSource);
        } catch (ClassNotFoundException e) {
            LOGGER.error("创建数据源失败", e);
            throw new RuntimeException(e);
        }
    }

添加前的SQL，添加之后的SQL日志


注意⚠️
    p6spy本身配合slf4j+log4j使用，项目中需要整合logback，目前小编的项目还有一点不完美就是，sql.log日志文件不能自动自动切割，如果一定要利用logback来完成，p6spy打印的所有日志，输出到main函数所在的目录下。如果一定希望能自动切割，那就多关注一些main函数所在的包的日志切割情况。

总结

    之前自己对技术的应用大部分停留到应用的层面，很片面的认为很少有机会去根据项目的需要改动现有框架的源码，所以会用就Ok了。但是这次在项目中整合p6spy，因为日志框架的问题，再加上对p6spy深入认知不够，还尚未完美的将p6spy+logback整合到一起。也让我反思对技术为什么要追求更高层次上的认知！