{"id":"https://openalex.org/W4400909796","doi":"https://doi.org/10.1109/icde60146.2024.00456","title":"Evaluating Text-to-SQL Model Failures on Real-World Data","display_name":"Evaluating Text-to-SQL Model Failures on Real-World Data","alternate_titles":[],"publication_year":2024,"publication_date":"2024-05-13","ids":{"openalex":"https://openalex.org/W4400909796","doi":"https://doi.org/10.1109/icde60146.2024.00456"},"language":"en","primary_location":{"id":"doi:10.1109/icde60146.2024.00456","doi":"https://doi.org/10.1109/icde60146.2024.00456","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icde60146.2024.00456","pdf_url":null,"source":{"id":"https://openalex.org/S4210210321","display_name":"IEEE International Conference on Data Engineering (ICDE)","issn_l":"1084-4627","issn":["1084-4627","2375-0286"],"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":["norway-2"],"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IEEE 40th International Conference on Data Engineering (ICDE)","raw_type":"proceedings-article"},"parent_works":[],"lineage":["https://openalex.org/W4400909796"],"child_works_count":0,"type":"conference-abstract","is_traditional":false,"indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5104169675","display_name":"Manasi Ganti","orcid":"https://orcid.org/0009-0000-8213-7286","observed_orcids":["https://orcid.org/0009-0000-8213-7286"]},"institutions":[{"id":"https://openalex.org/I201448701","display_name":"University of Washington","ror":"https://ror.org/00cvxb145","country_code":"US","type":"education","lineage":["https://openalex.org/I201448701"]}],"countries":["US"],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Manasi Ganti","raw_affiliation_strings":["University of Washington,Seattle"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Washington,Seattle","institution_ids":["https://openalex.org/I201448701"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5054285881","display_name":"Laurel Orr","orcid":"https://orcid.org/0000-0002-2183-3541","observed_orcids":["https://orcid.org/0000-0002-2183-3541"]},"institutions":[],"countries":[],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Laurel Orr","raw_affiliation_strings":["Numbers Station AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Numbers Station AI","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5033975637","display_name":"Sen Wu","orcid":"https://orcid.org/0000-0002-6133-4122","observed_orcids":["https://orcid.org/0000-0002-6133-4122"]},"institutions":[],"countries":[],"is_corresponding":false,"contributor_roles":[],"raw_author_name":"Sen Wu","raw_affiliation_strings":["Numbers Station AI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Numbers Station AI","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":15,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"1"},"is_retracted":false,"is_pulled":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.5390999913215637,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.5390999913215637,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10317","display_name":"Advanced Database Systems and Queries","score":0.21150000393390656,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.11100000143051147,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/text-to-sql","display_name":"text-to-SQL","score":0.9290000200271606},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.6669999957084656},{"id":"https://openalex.org/keywords/schema-complexity","display_name":"schema complexity","score":0.6830000281333923},{"id":"https://openalex.org/keywords/join-queries","display_name":"join queries","score":0.7929999828338623},{"id":"https://openalex.org/keywords/nested-queries","display_name":"nested queries","score":0.7289999723434448},{"id":"https://openalex.org/keywords/question-ambiguity","display_name":"question ambiguity","score":0.824999988079071},{"id":"https://openalex.org/keywords/spider-benchmark","display_name":"Spider benchmark","score":0.625},{"id":"https://openalex.org/keywords/gpt-4","display_name":"GPT-4","score":0.7670000195503235},{"id":"https://openalex.org/keywords/starcoder","display_name":"StarCoder","score":0.6460000276565552},{"id":"https://openalex.org/keywords/query-generation","display_name":"query generation","score":0.42399999499320984},{"id":"https://openalex.org/keywords/benchmark-design","display_name":"benchmark design","score":0.335999995470047},{"id":"https://openalex.org/keywords/execution-accuracy","display_name":"execution accuracy","score":0.5730000138282776}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7907915711402893},{"id":"https://openalex.org/C510870499","wikidata":"https://www.wikidata.org/wiki/Q47607","display_name":"SQL","level":2,"score":0.5844724178314209},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.4949924349784851},{"id":"https://openalex.org/C77088390","wikidata":"https://www.wikidata.org/wiki/Q8513","display_name":"Database","level":1,"score":0.45371541380882263},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.4489075243473053}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icde60146.2024.00456","doi":"https://doi.org/10.1109/icde60146.2024.00456","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icde60146.2024.00456","pdf_url":null,"source":{"id":"https://openalex.org/S4210210321","display_name":"IEEE International Conference on Data Engineering (ICDE)","issn_l":"1084-4627","issn":["1084-4627","2375-0286"],"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":["norway-2"],"host_organization":"https://openalex.org/P4310319808","host_organization_name":"Institute of Electrical and Electronics Engineers","host_organization_lineage":["https://openalex.org/P4310319808"],"host_organization_lineage_names":["Institute of Electrical and Electronics Engineers"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2024 IEEE 40th International Conference on Data Engineering (ICDE)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4403836156","https://openalex.org/W7133239790","https://openalex.org/W6926019341","https://openalex.org/W4399424215","https://openalex.org/W7166295915","https://openalex.org/W4389523781","https://openalex.org/W4402955398","https://openalex.org/W7078726768","https://openalex.org/W4410234067","https://openalex.org/W7168995906"],"abstract_inverted_index":{"Text-to-SQL":[0,81,110,210,269],"generation":[1],"models,":[2,34,211],"capable":[3],"of":[4,148,181,208,282,291,316,319,355,362,401],"converting":[5],"natural":[6],"language":[7,33],"prompts":[8],"into":[9],"SQL":[10],"queries,":[11,163,389],"offer":[12],"significant":[13],"potential":[14],"for":[15,70,230,253,261,380],"streamlining":[16],"data":[17],"analytics":[18],"tasks.":[19],"Despite":[20],"state-of-the-art":[21],"performance":[22,41,207,370,409],"on":[23,43,84,91,95,101,151,222,294,296,308],"popular":[24],"academic":[25,66,309,363,375],"benchmarks":[26,67,310,404],"such":[27,35],"as":[28,36],"Spider":[29,228,235],"[1],":[30],"recent":[31],"large":[32,182],"GPT-4,":[37,213],"exhibit":[38],"a":[39,127],"considerable":[40],"degradation":[42],"real-world":[44,72,109,283,300,411],"applications":[45],"with":[46,272,340],"longer,":[47],"more":[48],"convoluted":[49],"schemas":[50,347,358],"[2].":[51],"This":[52],"disparity":[53,371],"raises":[54],"questions":[55,297,304],"about":[56,257],"what":[57],"factors":[58],"contribute":[59],"to":[60,144,190,200,238,303,323,336,398,405],"this":[61,368],"drop":[62,315],"and":[63,118,136,141,160,218,226,276,311,321,374,390],"whether":[64],"existing":[65,138,268],"are":[68,280],"effective":[69],"representing":[71],"challenges.":[73],"To":[74,177],"determine":[75],"these":[76,123],"factors,":[77],"we":[78,125,184],"first":[79],"examine":[80],"model":[82,152,240,369],"failures":[83],"customer":[85,92,134],"logs.":[86],"We":[87,103,204,233,396],"find":[88],"that":[89,267,279,298,361],"accuracy":[90,100,237,252,290,314,350],"logs":[93,135],"was":[94],"average":[96,295,318],"30%":[97],"lower":[98],"than":[99],"Spider.":[102],"identify":[104],"three":[105,167],"main":[106],"challenges":[107],"in":[108,173,250,288,382,410],"applications:":[111],"long":[112,384],"context":[113,385],"length,":[114],"unclear":[115],"question":[116,169,394],"formulation,":[117],"greater":[119],"query":[120],"complexity.":[121],"With":[122],"insights,":[124],"create":[126],"new":[128],"benchmark":[129,155,225,229,366],"built":[130],"from":[131,188,360],"manually":[132],"labeled":[133],"evaluate":[137],"open":[139],"source":[140],"private":[142],"LLMs":[143],"demonstrate":[145],"the":[146,179,196,206,227,353,378],"impact":[147],"each":[149,164],"factor":[150],"performance.":[153,241],"The":[154,242],"incorporates":[156],"20":[157],"non-join":[158],"queries":[159,175,334,356],"30":[161],"join":[162],"accompanied":[165],"by":[166],"additional":[168,263],"phrasing":[170,301],"variations,":[171],"resulting":[172],"200":[174],"total.":[176],"capture":[178,407],"effects":[180],"schemas,":[183,255],"vary":[185],"schema":[186,277],"size":[187],"5":[189],"over":[191,330],"300":[192],"columns":[193,198],"while":[194],"retaining":[195],"minimum":[197],"required":[199],"answer":[201],"all":[202],"questions.":[203],"assess":[205],"prominent":[209],"including":[212],"GPT-3.5,":[214],"BigCode's":[215],"Starcoder":[216],"[3],":[217],"NSQL":[219],"Llama-2":[220],"[4]":[221],"both":[223],"our":[224],"comparative":[231],"analysis.":[232],"use":[234],"execution":[236,251,289],"measure":[239],"evaluation":[243],"results":[244],"reveal":[245],"a)":[246],"A":[247,286],"consistent":[248],"decline":[249],"longer":[254],"dropping":[256],"0.5":[258],"percentage":[259],"points":[260,293,325],"every":[262],"10":[264],"columns,":[265],"indicating":[266],"models":[270,327,341],"struggle":[271],"progressively":[273],"larger":[274],"tables":[275],"lengths":[278],"characteristic":[281],"datasets,":[284],"b)":[285],"decrease":[287],"12.3":[292],"emulate":[299],"compared":[302,335],"phrased":[305],"unambiguously":[306],"based":[307],"c)":[312],"An":[313],"an":[317],"36":[320],"up":[322],"52":[324],"when":[326,352],"must":[328],"reason":[329],"nested":[331],"or":[332,357],"complex":[333,388],"simple":[337],"SELECT":[338],"statements,":[339],"commonly":[342],"making":[343],"errors":[344],"interpreting":[345],"column":[346],"correctly.":[348],"Overall,":[349],"drops":[351],"complexity":[354],"increases":[359],"benchmarks.":[364],"Our":[365],"highlights":[367],"between":[372],"enterprise":[373,402],"settings,":[376],"emphasizing":[377],"need":[379],"improvements":[381],"handling":[383],"tasks,":[386],"generating":[387],"increasing":[391],"robustness":[392],"against":[393],"ambiguity.":[395],"hope":[397],"encourage":[399],"development":[400],"inspired":[403],"better":[406],"LLM":[408],"scenarios.":[412]},"cited_by_api_url":"https://api.openalex.org/works?filter=cites:W4400909796","counts_by_year":[{"year":2026,"cited_by_count":4},{"year":2025,"cited_by_count":7},{"year":2024,"cited_by_count":4}],"updated_date":"2026-11-21T10:27:49.728292","created_date":"2024-07-24T00:00:00"}
