@inproceedings{shahbazi-etal-2026-textual,
title = "From Textual Columns to Query Plans: A Unified Relational-Semantic Execution Framework for Hybrid Query Processing",
author = "Shahbazi, Nima and
Maekawa, Seiji and
Bhutani, Nikita and
Hruschka, Estevam",
editor = "Zhang, Yue and
Xu, Ruifeng and
Purwarianti, Ayu and
Rong, Tong and
Suzuki, Jun",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {IJCNLP}-{AACL} 2026",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-aacl.200/",
pages = "3979--4002",
ISBN = "979-8-89176-444-6",
abstract = "Existing table question answering systems face a key trade-off: approaches designed for relational databases scale efficiently but assume fully structured schemas, while LLM-based methods support flexible reasoning over unstructured text but struggle to scale to large and multi-table datasets. This limits their applicability to semi-structured data, where answering a query may require dynamically recovering latent schema elements from free-form text before symbolic execution can proceed. We present OmniTQA, an operator-centric framework that unifies relational execution and LLM-based semantic reasoning within a single query runtime. OmniTQA compiles queries into directed acyclic graphs (DAGs) of relational and semantic operators, enabling cost-aware optimization, dual-engine execution, and parallel exploration of alternative semantic groundings under ambiguity. Experiments on structured and semi-structured benchmarks show that OmniTQA substantially outperforms existing approaches on hybrid data settings while remaining competitive on structured datasets."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="shahbazi-etal-2026-textual">
<titleInfo>
<title>From Textual Columns to Query Plans: A Unified Relational-Semantic Execution Framework for Hybrid Query Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Nima</namePart>
<namePart type="family">Shahbazi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Seiji</namePart>
<namePart type="family">Maekawa</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nikita</namePart>
<namePart type="family">Bhutani</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Estevam</namePart>
<namePart type="family">Hruschka</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayu</namePart>
<namePart type="family">Purwarianti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Rong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-444-6</identifier>
</relatedItem>
<abstract>Existing table question answering systems face a key trade-off: approaches designed for relational databases scale efficiently but assume fully structured schemas, while LLM-based methods support flexible reasoning over unstructured text but struggle to scale to large and multi-table datasets. This limits their applicability to semi-structured data, where answering a query may require dynamically recovering latent schema elements from free-form text before symbolic execution can proceed. We present OmniTQA, an operator-centric framework that unifies relational execution and LLM-based semantic reasoning within a single query runtime. OmniTQA compiles queries into directed acyclic graphs (DAGs) of relational and semantic operators, enabling cost-aware optimization, dual-engine execution, and parallel exploration of alternative semantic groundings under ambiguity. Experiments on structured and semi-structured benchmarks show that OmniTQA substantially outperforms existing approaches on hybrid data settings while remaining competitive on structured datasets.</abstract>
<identifier type="citekey">shahbazi-etal-2026-textual</identifier>
<location>
<url>https://aclanthology.org/2026.findings-aacl.200/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>3979</start>
<end>4002</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T From Textual Columns to Query Plans: A Unified Relational-Semantic Execution Framework for Hybrid Query Processing
%A Shahbazi, Nima
%A Maekawa, Seiji
%A Bhutani, Nikita
%A Hruschka, Estevam
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-444-6
%F shahbazi-etal-2026-textual
%X Existing table question answering systems face a key trade-off: approaches designed for relational databases scale efficiently but assume fully structured schemas, while LLM-based methods support flexible reasoning over unstructured text but struggle to scale to large and multi-table datasets. This limits their applicability to semi-structured data, where answering a query may require dynamically recovering latent schema elements from free-form text before symbolic execution can proceed. We present OmniTQA, an operator-centric framework that unifies relational execution and LLM-based semantic reasoning within a single query runtime. OmniTQA compiles queries into directed acyclic graphs (DAGs) of relational and semantic operators, enabling cost-aware optimization, dual-engine execution, and parallel exploration of alternative semantic groundings under ambiguity. Experiments on structured and semi-structured benchmarks show that OmniTQA substantially outperforms existing approaches on hybrid data settings while remaining competitive on structured datasets.
%U https://aclanthology.org/2026.findings-aacl.200/
%P 3979-4002
Markdown (Informal)
[From Textual Columns to Query Plans: A Unified Relational-Semantic Execution Framework for Hybrid Query Processing](https://aclanthology.org/2026.findings-aacl.200/) (Shahbazi et al., Findings 2026)
ACL