linux - 将 bash 日志文件转换为 CSV

Question

我在 Linux 服务器上有一些应用程序日志数据，看起来像这样，

     2015-10-09 08:16:40,692 28456 INFO Demodb openerp.addons.service_logger.service_logger: 1 res.partner.fields_view_get ([], {'view_id': 127, 'toolbar': False, 'context': {'
     lang': 'en_US', 'tz': 'Europe/Brussels', 'uid': 1, 'active_model': 'res.partner', 'params': {'action': 60}, 'search_default_customer': 1}, 'view_type': 'search'}) 
     2015-10-09 08:16:40,701 28456 INFO Demodb openerp.addons.service_logger.service_logger: 1 res.partner.fields_view_get ([], {'view_id': 128, 'toolbar': True, 'context': {'l
     ang': 'en_US', 'tz': 'Europe/Brussels', 'uid': 1, 'active_model': 'res.partner', 'params': {'action': 60}, 'search_default_customer': 1}, 'view_type': 'kanban'}) 

我想把这个输出成csv格式，(基本上从'([]'可以是一个batch)

     2015-10-09,08:16:40,701,28456,INFO,Demodb openerp.addons.service_logger.service_logger:,1,res.partner.fields_view_get,([],{'view_id': 128,'toolbar': True,'context':,{'l
     ang':'en_US','tz': 'Europe/Brussels', 'uid': 1, 'active_model': 'res.partner', 'params': {'action': 60}, 'search_default_customer': 1}, 'view_type': 'kanban'}) 

任何输入都会有帮助，

Answer 1

这会将所有空格替换为逗号，直到第一次出现 ([]:

sed -e 'h;s/.*(\[\]//;x;s/\s\+(\[\].*//g;s/\s\+/,/g;G;s/\n/,([]/'

解释:

h                 # copy line to hold
s/.*(\[\]//       # remove everything before ([]
x                 # swap unchanged line back from hold, put part after ([] to hold
s/\s\+(\[\].*//g  # remove everything after ([]
s/\s\+/,/g        # do work, that is change spaces to commas
G                 # glue the two parts together, with a newline inbetween
s/\n/,([]/        # put the ([]-marker back between the two parts