http://tomasrohr.org/wiki/ 2026-05-03T11:19:10+02:00 http://tomasrohr.org/wiki/ http://tomasrohr.org/wiki/lib/tpl/mnml-blog/images/favicon.ico text/html 2017-11-22T13:11:56+02:00 http://tomasrohr.org/wiki/doku.php?id=blog:apache_kafka_tips_1_-_topics&rev=1511352716&do=diff Apache Kafka Tips #1 - Topics 1) Pojmenování topiků Název musí popisovat data uvnitř. Jeden topik = jedna entita datového modelu (zákazník, metriky z měřidel z daného okamžiku, zařízení, lokace) Též topik = jeden druh logických událostí (naměřené hodnoty z měřidel, události čidel (otevřené okno, atd), alerty (např. čidlo mimo provoz, otevření dveří po půlnoci) ) text/html 2019-03-29T14:03:17+02:00 http://tomasrohr.org/wiki/doku.php?id=blog:incremental_insert_in_hive_and_impala&rev=1553864597&do=diff Incremental insert in Hive and Impala Task: To insert data set into a table, but if a record already exists in the target table, it must not be inserted. And you must overwrite whole partitions (which is a good practice on Hive). The source data set may be a table or a query result. text/html 2019-03-29T14:02:40+02:00 http://tomasrohr.org/wiki/doku.php?id=blog:merge_in_hive&rev=1553864560&do=diff Merge in Hive and Impala Suppose you want to load data into an already populated Hive table and you want to apply data as update of existing records and insert of new records. Neither Hive nor Impala query language includes MERGE command (or UPSERT known from Teradata). Also it is a good practice to ovewrite a whole table partition at once. Here is the command which performs what you would achieve in RDBMS using