使用CSV格式的非結構化GPS數據包創建結構化配置單元表

Question

我有一個如下所述的csv文件。

VTS,51,0071,9739965515,NM,GP,INF01,V,19,072219,291014,0000.0000,N,00000.0000,E,07AE VTS,01,0097,9739965515,SP,GP,18,072253,V,0000.0000,N,00000.0000,E,0.0,0.0,291014,0000,00,4000,11,999,169,B205 VTS,51,0071,9739965515,NM,GP,INF01,V,18,072311,291014,0000.0000,N,00000.0000,E,C24E VTS,01,0097,9739965515,NM,GP,19,072311,V,0000.0000,N,00000.0000,E,0.0,0.0,291014,0000,00,4000,11,999,171,B358 VTS,51,0071,9739965515,NM,GP,INF01,V,18,072319,291014,0000.0000,N,00000.0000,E,012F VTS,51,0071,9739965515,NM,GP,INF01,V,19,072326,291014,0000.0000,N,00000.0000,E,B2E6 VTS,01,0097,9739965515,NM,GP,18,072326,V,0000.0000,N,00000.0000,E,0.0,0.0,291014,0000,00,4000,11,999,173,EAA0 VTS,51,0071,9739965515,NM,GP,INF01,V,18,072333,291014,0000.0000,N,00000.0000,E,9896 VTS,51,0071,9739965515,NM,GP,INF01,V,18,072340,291014,0000.0000,N,00000.0000,E,9B23

這必須與字段映射：

pkt_header,gprs_pkt_id,pkt_length,sim_no,msg_id,gprs_pkt,gsm_sig_strength,utc_time,pkt_validation,latitude,direction_n_s,longitude,direction_e_w,speed,track_angle,utc_date,fuel_adc_values,ignition,odometer_values,supply_int,battery_adc,pkt_id,check_sum

第二個字段，即值為01的gprs_pkt_id描述了一個有效的數據包。 我的用例是僅針對有效數據包過濾CSV數據，而我正在使用正則表達式，但是我無法獲取全部數據。 任何幫助將不勝感激。

使用的Hive查詢如下所示。

CREATE EXTERNAL TABlE sky_track_testing1( pkt_header STRING, gprs_pkt_id STRING, pkt_length STRING, sim_no STRING, msg_id STRING, gprs_pkt STRING, gsm_sig_strength STRING, utc_time STRING, pkt_validation STRING, latitude STRING, direction_n_s STRING, longitude STRING, direction_e_w STRING, speed STRING, track_angle STRING, utc_date STRING, fuel_adc_values STRING, ignition STRING, odometer_values STRING, supply_int STRING, battery_adc STRING, pkt_id STRING, check_sum STRING ) ROW FORMAT SERDE 'org.apache.hadoop.hive.contrib.serde2.RegexSerDe' WITH SERDEPROPERTIES ( "input.regex" = "^(VTS,01).*$" ) STORED AS TEXTFILE LOCATION '/user/root/sky_track';

這絕對是錯誤的查詢。 請幫我。

Answer 1

我建議您為此使用Pig ：

a = load '/user/root/sky_track' as (pkt_header,gprs_pkt_id,pkt_length,sim_no,msg_id,gprs_pkt,gsm_sig_strength,utc_time,pkt_validation,latitude,direction_n_s,longitude,direction_e_w,speed,track_angle,utc_date,fuel_adc_values,ignition,odometer_values,supply_int,battery_adc,pkt_id,check_sum);
b = filter a by gprs_pkt_id == '01';
store b into '/user/root/sky_track_valid';

Answer 2

是的，根據以上答案，Pig非常適合您的數據。 你可以試試豬。 如果您對配置單元感興趣，請參見以下示例（數據集不需要正則表達式）。

hive> CREATE  TABLE sky_track_testing1(
    > pkt_header STRING,
    > gprs_pkt_id STRING,
    > pkt_length STRING,
    > sim_no STRING,
    > msg_id STRING,
    > gprs_pkt STRING,
    > gsm_sig_strength STRING,
    > utc_time STRING,
    > pkt_validation STRING,
    > latitude STRING,
    > direction_n_s  STRING,
    > longitude  STRING,
    > direction_e_w STRING,
    > speed STRING,
    > track_angle  STRING,
    > utc_date STRING,
    > fuel_adc_values STRING,
    > ignition  STRING,
    > odometer_values STRING,
    > supply_int  STRING,
    > battery_adc  STRING,
    > pkt_id  STRING,
    > check_sum STRING
    > ) 
    > ROW FORMAT
    > DELIMITED FIELDS TERMINATED BY ','
    > LINES TERMINATED BY '\n'
    > STORED AS TEXTFILE;
OK
Time taken: 0.1 seconds

hive> select *from sky_track_testing1 where gprs_pkt_id='01';
OK
VTS 01  0097    9739965515  SP  GP  18  072253  V   0000.0000   N   00000.0000  E   0.0 0.0 291014  0000    00  4000    1999    169 B205
VTS 01  0097    9739965515  NM  GP  19  072311  V   0000.0000   N   00000.0000  E   0.0 0.0 291014  0000    00  4000    1999    171 B358
VTS 01  0097    9739965515  NM  GP  18  072326  V   0000.0000   N   00000.0000  E   0.0 0.0 291014  0000    00  4000    1999    173 EAA0
Time taken: 14.328 seconds