Alvaro te copio el resultado acá: id 101 subclient_id 101 100 hidden_by_contact 101 6 fulldate 101 title 101 33 pointers 101 100 page_num 101 100 eurl 11 url 101 1 ep_file 101 epoch_time 101 36 add_type 10 15 alt_title 101 10 views 12 6 box 8 23 vp 101 100 clip_ratio 101 2 clip_area 101 tirada 101 75 difusion 101 99 audiencia 101 100 file_type 3 score 101 33 meta_description 101 1 tono 3 saved 2 sortnum 39 domain_id 101 100 approved 1 manual_domains_id duplicated 2 created_pdf_type 4 7 pdf_created_by 22 52 vo 5 url_views 6 5 print_section_id advertisement 2 pdf_verified num_pages 21 12 available 1 evp 101 100 com_value 101 100 companies_group_id 101 is_ungrouped 2 sentiment_score 11 p_vistas 101 100 main_tema_id 101 100 secondary_tema_ids 101 1 enable_eng_tracking 1 engagement_count 5 1 eng_last_tracked 69 usid 101 100 unique_article_id 101 emailed_at emailed_by_instance_id last_change_date 101 1 tracked_companies_id 101 paid_article_grab 2 paid_article_conv 2 processing_fulldate 101 1 pdf_downloads custom_fields_values
El mié, 5 feb 2025 a las 12:49, Álvaro Herrera (<alvhe...@alvh.no-ip.org>) escribió: > Guillermo E. Villanueva escribió: > > Hola Alvaro cómo estas? > > Los índices de los que hablo tienen unos 27GB cada uno > > La tabla tiene unos 1.4TB > > > > effective_cache_size = 48GB > > shared_buffers = 24GB > > ¿y qué dice esto? > > select attname, cardinality(histogram_bounds), > cardinality(most_common_freqs) > from pg_stats > where tablename = 'companies'; > > -- > Álvaro Herrera PostgreSQL Developer — > https://www.EnterpriseDB.com/ >
