Log processing 3
Task:Each piece of log has indefinite rows, and a fixed mark is at the beginning of each record.
Python
| 1 | import pandas as pd |
| 2 | log_file = ‘E://txt//Indefinite _info2.txt‘ |
| 3 | log_info = pd.read_csv(log_file,header=None) |
| 4 | group_cond = log_info[0].apply(lambda x:1 if x.split("\t")[0].split(":")[0]=="userid" else 0).cumsum() |
| 5 | log_g = log_info.groupby(group_cond,sort=False) |
| 6 | columns = ["userid","gender","age","salary","province","musicid","watch_time","time"] |
| 7 | df_dic = {} |
| 8 | for c in columns: |
| 9 | df_dic[c]=[] |
| 10 | for index,group in log_g: |
| 11 | rec_dic = {} |
| 12 | rec = group.values.flatten() |
| 13 | rec = ‘\t’.join(rec).split("\t") |
| 14 | for r in rec: |
| 15 | v = r.split(":") |
| 16 | rec_dic[v[0]]=v[1] |
| 17 | for col in columns: |
| 18 | if col not in rec_dic.keys(): |
| 19 | df_dic[col].append(None) |
| 20 | else: |
| 21 | df_dic[col].append(rec_dic[col]) |
| 22 | df = pd.DataFrame(df_dic) |
| 23 | print(df) |
Pandas does not have the function of grouping by conditions. It needs to construct an array of grouping by conditions.
esProc
| A | ||
| 1 | E://txt//Indefinite _info2.txt | |
| 2 | [userid,gender,age,salary,province,musicid,watch_time,time] | |
| 3 | =file(A1).import@s() | |
| 4 | =A3.group@i(_1.array("\t")(1).array("\:")(1)=="userid") | |
| 5 | =A4.(~.(_1.array("\t")).conj().align(A2,~.array("\:")(1)).(~.array("\:")(2))).conj() | |
| 6 | =create(${A2.concat@c()}).record(A5) |
esProc has powerful grouping function and loop computing ability, and its code is simple and clear.