我有一个难题，我想从一个有特殊字的文本中删除一些块，这里是一个例子，我想要的是删除匹配backup_all和匹配行前的一行，以及匹配行后的三行< / p>
# Time: 2018-01-23T03:41:41.454104+08:00
# User@Host: backup_all[backup_all] @ [127.0.0.1] Id: 3168695
# Query_time: 0.129250 Lock_time: 0.000062 Rows_sent: 3535 Rows_examined: 3535
SET timestamp=1516650101;
SELECT /*!40001 SQL_NO_CACHE */ * FROM `legend_gather_customer_note_effect`;
# Time: 2018-01-23T03:41:41.527587+08:00
# User@Host: backup_all[backup_all] @ [127.0.0.1] Id: 3168695
# Query_time: 0.066378 Lock_time: 0.000059 Rows_sent: 193 Rows_examined: 193
SET timestamp=1516650101;
SELECT /*!40001 SQL_NO_CACHE */ * FROM `legend_gather_performance_config`;
# Time: 2018-01-23T03:41:41.558254+08:00
# User@Host: backup_all[backup_all] @ [127.0.0.1] Id: 3168695
# Query_time: 0.025533 Lock_time: 0.000058 Rows_sent: 296 Rows_examined: 296
SET timestamp=1516650101;
SELECT /*!40001 SQL_NO_CACHE */ * FROM `legend_gift`;
# Time: 2018-01-23T03:42:09.536559+08:00
# User@Host: zabbix_agent[zabbix_agent] @ [127.0.0.1] Id: 3169056
# Query_time: 0.000304 Lock_time: 0.000162 Rows_sent: 1 Rows_examined: 1
SET timestamp=1516650129;
SELECT SUM(trx_rows_locked) AS rows_locked, SUM(trx_rows_modified) AS rows_modified, SUM(trx_lock_memory_bytes) AS lock_memory FROM information_schema.INNODB_TRX;
删除后，文本成为以下输出，如何使用sed来实现这个目标？
# Time: 2018-01-23T03:42:09.536559+08:00
# User@Host: zabbix_agent[zabbix_agent] @ [127.0.0.1] Id: 3169056
# Query_time: 0.000304 Lock_time: 0.000162 Rows_sent: 1 Rows_examined: 1
SET timestamp=1516650129;
SELECT SUM(trx_rows_locked) AS rows_locked, SUM(trx_rows_modified) AS rows_modified, SUM(trx_lock_memory_bytes) AS lock_memory FROM information_schema.INNODB_TRX;
更新
这个文本来自mysql慢日志，我想从用户'backup_all'中删除所有sql信息，如你所见，用户'backup_all'的这个文本产品肯定是常规的，但是其他用户的其他文本产品可能不一样，他们会
# Time: 2018-01-23T03:41:24.490723+08:00
# User@Host: zabbix_agent[zabbix_agent] @ [127.0.0.1] Id: 3169038
# Query_time: 0.000669 Lock_time: 0.000334 Rows_sent: 1 Rows_examined: 27
SET timestamp=1516650084;
select count(*) Slownum from information_schema.processlist where COMMAND = 'Query' and info not like '%information_schema.processlist%' and TIME > 0;
# Time: 2018-01-23T03:41:40.628284+08:00
# User@Host: backup_all[backup_all] @ [127.0.0.1] Id: 3168695
# Query_time: 78.333179 Lock_time: 0.000073 Rows_sent: 13269064 Rows_examined: 13269064
SET timestamp=1516650100;
SELECT /*!40001 SQL_NO_CACHE */ * FROM `legend_final_inventory`;
# Time: 2018-01-23T03:41:40.925596+08:00
# User@Host: backup_all[backup_all] @ [127.0.0.1] Id: 3168695
# Query_time: 0.175956 Lock_time: 0.000065 Rows_sent: 5101 Rows_examined: 5101
SET timestamp=1516650100;
SELECT /*!40001 SQL_NO_CACHE */ * FROM `legend_finance_account`;
或
# Time: 2018-01-23T04:29:26.903048+08:00
# User@Host: yun_rw_legend[yun_rw_legend] @ [10.162.86.162] Id: 3167670
# Query_time: 0.000150 Lock_time: 0.000053 Rows_sent: 6 Rows_examined: 32
SET timestamp=1516652966;
select id, value from legend_precheck_value where value_type = 8 and is_deleted = 'N';
# Time: 2018-01-23T04:29:31.825823+08:00
# User@Host: yun_rw_legend[yun_rw_legend] @ [10.162.86.162] Id: 3167670
# Query_time: 0.000826 Lock_time: 0.000146 Rows_sent: 0 Rows_examined: 947
SET timestamp=1516652971;
select
id as id,
is_deleted as isDeleted,
gmt_create as gmtCreate,
creator as creator,
gmt_modified as gmtModified,
modifier as modifier,
shop_id as shopId,
customer_id as customerId,
account_id as accountId,
customer_car_id as customerCarId
from legend_customer_car_rel
WHERE is_deleted = 'N'
and shop_id = 3374
and customer_car_id = 1307177;
# Time: 2018-01-23T04:30:01.149529+08:00
# User@Host: backup_all[backup_all] @ [127.0.0.1] Id: 3170398
# Query_time: 0.003047 Lock_time: 0.000169 Rows_sent: 0 Rows_examined: 385
SET timestamp=1516653001;
SELECT LOGFILE_GROUP_NAME, FILE_NAME, TOTAL_EXTENTS, INITIAL_SIZE, ENGINE, EXTRA FROM INFORMATION_SCHEMA.FILES WHERE FILE_TYPE = 'UNDO LOG' AND FILE_NAME IS NOT NULL AND LOGFILE_GROUP_NAME IS NOT NULL GROUP BY LOGFILE_GROUP_NAME, FILE_NAME, ENGINE, TOTAL_EXTENTS, INITIAL_SIZE, EXTRA ORDER BY LOGFILE_GROUP_NAME;
# Time: 2018-01-23T04:30:01.151783+08:00
# User@Host: backup_all[backup_all] @ [127.0.0.1] Id: 3170398
# Query_time: 0.002082 Lock_time: 0.000119 Rows_sent: 0 Rows_examined: 385
SET timestamp=1516653001;
SELECT DISTINCT TABLESPACE_NAME, FILE_NAME, LOGFILE_GROUP_NAME, EXTENT_SIZE, INITIAL_SIZE, ENGINE FROM INFORMATION_SCHEMA.FILES WHERE FILE_TYPE = 'DATAFILE' ORDER BY TABLESPACE_NAME, LOGFILE_GROUP_NAME;
答案 0 :(得分：1)
似乎有效的强制性awk解决方案：
awk '
$2!~/backup_all/ {
printf "# Time%s", $0
}' RS='# Time' ORS='' FS='
' example_in
它将输入拆分为每个# Time的记录和\n的字段，以将每行作为字段。如果第二行（字段）中没有backup_all，print
第一个例子的输出：
# Time: 2018-01-23T03:42:09.536559+08:00
# User@Host: zabbix_agent[zabbix_agent] @ [127.0.0.1] Id: 3169056
# Query_time: 0.000304 Lock_time: 0.000162 Rows_sent: 1 Rows_examined: 1
SET timestamp=1516650129;
SELECT SUM(trx_rows_locked) AS rows_locked, SUM(trx_rows_modified) AS rows_modified, SUM(trx_lock_memory_bytes) AS lock_memory FROM information_schema.INNODB_TRX
和第二个：
# Time: 2018-01-23T04:29:26.903048+08:00
# User@Host: yun_rw_legend[yun_rw_legend] @ [10.162.86.162] Id: 3167670
# Query_time: 0.000150 Lock_time: 0.000053 Rows_sent: 6 Rows_examined: 32
SET timestamp=1516652966;
select id, value from legend_precheck_value where value_type = 8 and is_deleted = N;
# Time: 2018-01-23T04:29:31.825823+08:00
# User@Host: yun_rw_legend[yun_rw_legend] @ [10.162.86.162] Id: 3167670
# Query_time: 0.000826 Lock_time: 0.000146 Rows_sent: 0 Rows_examined: 947
SET timestamp=1516652971;
select
id as id,
is_deleted as isDeleted,
gmt_create as gmtCreate,
creator as creator,
gmt_modified as gmtModified,
modifier as modifier,
shop_id as shopId,
customer_id as customerId,
account_id as accountId,
customer_car_id as customerCarId
from legend_customer_car_rel
WHERE is_deleted = N
and shop_id = 3374
and customer_car_id = 1307177;
答案 1 :(得分：0)
这可能适合你（GNU sed）：
sed -n '/^#/{:a;N;/^SELECT/M!ba;/backup_all/!p}' file
由于这是对行的缩减/过滤，因此通过设置-n选项可以实现类似于grep的性质。收集以评论行（#）开头的行，并以以SELECT开头的行结束。如果这些行不包含单词backup_all，则打印它们。
N.B。 regexp上的M标志调用多行模式，其中^和$可以在行的开头/结尾匹配。另一种正则表达式是/\nSELECT/。