Question

我有一个Mnesia表格，其中包含以下格式：

-record(state, {key, tuple, state, timestamp, fin_from}).

条目如下（使用ets:tab2list(Tab)读取）：

[{state,{80,43252,tcp,tcp_syn_received,{192,168,101,5},{192,168,101,89}},
        {80,43252,tcp,{192,168,101,5},{192,168,101,89}},
        tcp_syn_received,1463850419221,undefined},
 {state,{80,41570,tcp,tcp_syn_received,{192,168,101,5},{192,168,101,89}},
        {80,41570,tcp,{192,168,101,5},{192,168,101,89}},
        tcp_syn_received,1463850403214,undefined},
...]

我想将这些数据写入.csv文件，每行一个条目 - 首选格式如下：

state,80,43252,tcp,tcp_syn_received,192.168.101.5,192.168.101.89,80,43252,tcp,192.168.101.5,192.168.101.89,tcp_syn_received,1463850419221,undefined
state,80,41570,tcp,tcp_syn_received,192.168.101.5,192.168.101.89,80,41570,tcp,192.168.101.5,192.168.101.89,tcp_syn_received,1463850419221,undefined

undefined之后应该有一个换行符。我尝试使用以下代码（而Content = ets:tab2list(states)）：

do_logging_async(File, Format, Content, Append)->
F = fun() ->
    file:write_file(File, io_lib:fwrite(Format, [Content]), [Append])
end,
spawn(F).

然而，我无法得到与我的输出相似的任何内容。之后应使用R评估数据。

更新：关键是逐行阅读表格并使用~w解析，但不是~p。我最终得到了以下解决方案（产生稍微不同的输出，但是，冗余数据更少）：

do_state_logging(File, EtsAsList) ->
% write header (columnnames)
file:write_file(File, io_lib:fwrite("~w,~w,~w,~w,~w,~w,~w,~w~n", [record,dstPort,srcPort,proto,dstIP,srcIP,state,timestamp]),[append]),
case EtsAsList of
    [] ->
        ok;
    _ ->
        F = fun({Record,_Key, 
            [P1, P2, Proto, {D_Ip_1,D_Ip_2,D_Ip_3,D_Ip_4}, {S_Ip_1,S_Ip_2,S_Ip_3,S_Ip_4}],
            State, Timestamp, _}) -> 
            file:write_file(File, io_lib:fwrite("~w,", [Record]),[append]),
            file:write_file(File, io_lib:fwrite("~w,~w,~w,", [P1,P2,Proto]),[append]),
            file:write_file(File, io_lib:fwrite("~w.~w.~w.~w,", [D_Ip_1,D_Ip_2,D_Ip_3,D_Ip_4]), [append]),
            file:write_file(File, io_lib:fwrite("~w.~w.~w.~w,", [S_Ip_1,S_Ip_2,S_Ip_3,S_Ip_4]), [append]),
            file:write_file(File, io_lib:fwrite("~w,", [State]),[append]),
            file:write_file(File, io_lib:fwrite("~w", [Timestamp]),[append]),
            file:write_file(File, ["\n"],[append]) 
        end,
        lists:foreach(F, EtsAsList)
    end,
    io:format("Finished logging of statetable to file: ~p~n" , [File]).

感谢回答谁让我想到了这个想法。

Answer 1

假设您将ets记录值更改为列表而不是元组，则可以使用此代码将ETS表写入文件。

do_logging_async(File, EtsAsList) ->
    F = fun({Key, Value}) -> 
        file:write_file(File, [atom_to_list(Key) ++ ","],[append]),
        write_value(File,lists:flatten(Value)),
        file:write_file(File, ["\n"],[append]) 
    end,
    lists:foreach(F,EtsAsList).


write_value(_File, []) -> ok;
write_value(File, [H|T]) ->
    case is_integer(H) of
        true -> file:write_file(File, [integer_to_list(H)],[append]);
        false -> file:write_file(File, [atom_to_list(H)],[append])
    end,
    case T=:=[] of
        true -> ok;
        false -> file:write_file(File, [","],[append])
    end,
    write_value(File,T).

do_logging_async/2需要每{Key, Value}对。首先，它将Key写入文件，然后在write_value/2上运行Value，在每个对的末尾写\n。

write_value/2获取展平值列表（假设它是一个仅包含整数和原子的展平列表）并将其写入文件。

Erlang：将mnesia表写入.csv每行一条记录

1 个答案: