1. Via Hive View
CREATE EXTERNAL TABLE if not exists finance.json_serde_optd_table (RetCode string, retmsg string, data array<struct < secid:string,= "" tradedate:date,= "" optid:string,= "" ticker:string,= "" secshortname:string,= "" EXCHANGECD: string,= "" presettleprice:double,= "" precloseprice:double,= "" openprice:double,= "" highestprice:double,= "" lowestprice:double,= "" closeprice:double,= "" settlprice:double,= "" turnovervol:double,= "" turnovervalue:double,= "" Openint:int= "" >>) ROW FORMAT SERDE ' org.apache.hive.hcatalog.data.JsonSerDe ' location ' hdfs://wdp.xxxxx.cn : 8020/nifi/finance1/optd/'; CREATE table if not exists Finance.tb_optdasselect b.data.secid,b.data.tradedate, B.data.optid,b.data.ticker,b.data.secshortname,b.data.exchangecd,b.data.presettleprice,b.data.precloseprice, B.data.openprice,b.data.highestprice,b.data.lowestprice,b.data.closeprice,b.data.settlprice,b.data.turnovervol , b.data.turnovervalue,b.data.openintfrom finance.json_serde_optd_table lateral VIEW explode (json_serde_optd_tAble.data) b as data;
?
2. By Zeppelin
?
%depz.load ("/usr/hdp/2.4.2.0-258/hive-hcatalog/share/hcatalog/hive-hcatalog-core.jar");
?
Define a collection of imported hive objects case Class Hiveconfig (database:string, modelname:string, hdfspath:string, schema:string, schema_tb:st ring); var hiveconfiglist = List[hiveconfig] ();
?
Create EQUD data structure//define JSON structure val Schema_json_equd_serde = "" "RetCode string, retmsg string, Data array<struct< secid= "": = "" string,= "" tradedate= "" date,= "" ticker= "" Secshortname= "" exchangecd= "precloseprice=" "double,=" "actprecloseprice:=" "openprice=" "highestprice=" "lowestprice=" "Closeprice = "Turnovervol=" "turnovervalue=" "dealamount=" "int,=" "turnoverrate=" "accumadjfactor=" "negmarketvalue=" " Marketvalue= "pe=" "pe1=" "pb=" "isopen=" "int=" ">>" "; var schema_equd =" "B.data.secid, B.data.ticker, B.data.secshortname, B.DATA.EXCHANGECD, B.data.tradedate, B.data.precloseprice , B.data.actprecloseprice, B.data.openprice, B.data.highestprice, B.data.lowes Tprice, B.data.closeprice, B.data.turnovervol, B.data.turnovervalue, b.data.de Alamount, B.data.turnoverrate, B.data.accumadjfactor, B.data.negmarketvalue, B.data.marketvalue, b.data.pe, B.data.pe1, B.DATA.PB, B.data.isopen "" "; hiveconfiglist = hiveconfiglist: + Hiveconfig (" F Inance "," Equd "," hdfs://wdp.xxxxx.cn:8020/nifi/finance1/", Schema_json_equd_serde, Schema_equd);
?
Create IDXD data structure//define JSON structure val Schema_json_idxd_serde = "" "RetCode string, retmsg string, Data array<struct< indexid:string,= "" tradedate:date,= "" ticker:string,= "" Porgfullname:stri ng,= "" secshortname:string,= "" exchangecd:string,= "" precloseindex:double,= "" openindex:double,= "" Lowestindex: double,= "highestindex:double,=" "closeindex:double,=" "turnovervol:double,=" "turnovervalue:double,=" "chg:double , = "chgpct:double=" ">>" "; var schema_idxd =" "B.data.indexid, B.data.tradedate, B.data.ticker , B.data.porgfullname, B.data.secshortname, B.DATA.EXCHANGECD, B.data.preclose Index, B.data.openindex, B.data.lowestindex, B.data.highestindex, B.data.close Index, B.data.turnovervol, B.data.turnovervalue, B.data.chg, b.data.chgpct ""; Hiveconfiglist = hiveconfiglist: + HiveconFig ("Finance", "Idxd", "hdfs://wdp.xxxxx.cn:8020/nifi/finance1/", Schema_json_idxd_serde, SCHEMA_IDXD);
?
Cyclic loading data in Def loaddatatohive (Args:hiveconfig) {val LoadPath = Args.hdfspath + args.modelname; Val Tb_json_serde = "Json_serde_" + args.modelname + "_table"; Val tb= "Tb_" + args.modelname; Val hivecontext = new Org.apache.spark.sql.hive.HiveContext (SC) if (args.database! = "" && args.schema! = "") { Print ("Creating project ..." + args.modelname) hivecontext.sql ("Create DATABASE IF not EXISTS" + args.database); Print ("Building extension model ..."); Hivecontext.sql ("CREATE TABLE IF not EXISTS" + Args.database + "." + Tb_json_serde + "(" + Args.schema + ") row format SE Rde ' org.apache.hive.hcatalog.data.JsonSerDe ' location "+" ' "+ LoadPath +"/"); println ("CREATE TABLE IF not EXISTS" + Args.database + "." + TB + "AS-select" + ARGS.SCHEMA_TB + "from" + Args.databa Se + "." + Tb_json_serde + "lateral VIEW explode (" + Tb_json_serde + ". Data) b as data"); Hivecontext.sql ("CREATE TABLE IF not EXISTS" + Args.database + "." + TB + "as select" +ARGS.SCHEMA_TB + "from" + Args.database + "." + Tb_json_serde + "lateral VIEW explode (" + Tb_json_serde + ". Data) B as Data "); println (Args.modelname + "extended model loading completed!"); }} hiveconfiglist.size; Hiveconfiglist.foreach {x = loaddatatohive (x)};
?
? 3. The second type of extraction
Since data is an array in the JSON data, the conversion above is a bit more complicated. The following approach is to first take the data array from the JSON and put it in HDFs, and then put the following statement directly into hive:
Use Splitjson to extract and delimit data arrays
CREATE EXTERNAL TABLE if not exists finance.awen_optd ( secid string, tradedate date, Optid string, Ticker string, secshortname string, exchangecd string, presettleprice double, precloseprice double, Openprice Double, highestprice double, lowestprice double, closeprice double, settlprice Double, Turnovervol double, turnovervalue double, openint int) ROW FORMAT SERDE ' Org.apache.hive.hcatalog.data.JsonSerDe ' location ' hdfs://wdp.xxxx.cn:8020/nifi/finance2/optd/';
?
Importing HDFs data to Hive