From 407f140f10450b817793275f7b10e9a6023d3900 Mon Sep 17 00:00:00 2001 From: James McKinney <26463+jpmckinney@users.noreply.github.com> Date: Wed, 1 May 2024 01:47:44 -0400 Subject: [PATCH] kingfisher-collect(incremental): Add transaction to replace _result table --- salt/kingfisher/collect/files/bi/cron.sh | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/salt/kingfisher/collect/files/bi/cron.sh b/salt/kingfisher/collect/files/bi/cron.sh index fb669bd9..afe89c0f 100644 --- a/salt/kingfisher/collect/files/bi/cron.sh +++ b/salt/kingfisher/collect/files/bi/cron.sh @@ -13,7 +13,7 @@ env SENTRY_DSN={{ SENTRY_DSN }} .ve/bin/scrapy crawl \ --logfile="{{ userdir }}/logs/{{ crawl.spider }}-$(date +%F).log" # shellcheck disable=all -{%- if crawl.powerbi %} +{%- if 'powerbi' in crawl and crawl.powerbi %} psql -U kingfisher_collect -h localhost -t -c 'SELECT data FROM {{ crawl.spider }}' -o {{ scratchdir }}/{{ crawl.spider }}.jsonl ocdscardinal prepare \ @@ -32,6 +32,9 @@ ocdscardinal indicators \ {{ userdir }}/bin/manage.py json-to-csv {{ scratchdir }}/{{ crawl.spider }}.json {{ scratchdir }}/{{ crawl.spider }}.csv psql postgresql://kingfisher_collect@localhost:5432/kingfisher_collect \ + -c "BEGIN" \ + -c "DELETE FROM {{ crawl.spider }}_result" \ -c "\copy {{ crawl.spider }}_result (ocid, subject, code, result, buyer_id, procuring_entity_id, tenderer_id, created_at) FROM stdin DELIMITER ',' CSV HEADER" \ + -c "END" \ < {{scratchdir}}/{{crawl.spider}}.csv {%- endif %}