What is “prefetching” in Odoo’s ORM and why does it matter?

Prefetching loads fields for an entire recordset at once, reducing repeated SQL queries. It prevents the N+1 problem and keeps large dataset operations fast.

When does prefetching fail and what breaks it?

Prefetching fails when records are browsed one-by-one or when loops repeatedly call search/browse. Splitting recordsets into single IDs removes Odoo’s ability to batch field loading.

When should I use search_read() instead of search() + read()?

Use search_read() when you need only a few fields for many records. It merges search and read into one call, reducing roundtrips and improving performance for list and API outputs.

What’s the efficient way to batch create or update records?

Create in bulk using lists or @api.model_create_multi, and update using a single write() on a recordset. This avoids repeated flushes and speeds up heavy computed-field operations.

When should I disable or customize prefetching manually?

Disable it when only minimal fields are needed from large datasets or when prefetching wastes memory. It helps in heavy cron jobs, imports, or cases where default prefetching adds overhead.

← Back to Blogs

Performance-Smart API Patterns in Odoo 19

Introduction

Bank reconciliation has always been a heavyweight chore for finance teams: repetitive matching, hunt-and-click corrections, and a mountain of exceptions that show up right when month-end pressure peaks.

Odoo 19 takes a big step forward by making reconciliation smarter and more action-oriented not by replacing judgement, but by reducing the routine work that eats up time and attention. Below I explain what changed, why it matters, and how to implement these improvements so your team gains real time back every close cycle.

What’s new in Odoo 19 — the highlights you’ll actually use

Odoo 19 brings several practical changes to the bank reconciliation workflow:

- Improved automatic reconciliation models: Rules that can run automatically are more accurate at recognizing recurring patterns and generating the correct counterpart entries. This reduces manual matching and repetitive write-offs.

- Reconcile draft entries: You can now reconcile draft journal entries; any automatic moves created (currency adjustments, cash-basis moves) are created in draft and the reconciliation is confirmed when the original is posted. This helps teams reconcile earlier in the process without waiting for everything to be posted.

- Better matching controls: Matching conditions include journal, partner, amount ranges, regex on labels, and more — so models can be very precise when they apply. When several models match, the sequence order decides which one runs first.

- Practical UI and keyboard help: The reconciliation view has keyboard shortcuts and preview features (like PDF preview for attached statements), speeding manual confirmations.

- Automatic runs and scheduled reconciliation: Automatic reconciliation can run on import, bank sync, or on scheduled jobs so many statement lines are matched without any daily clicking.

These changes are not flashy; they’re functional and focused on reducing clicks and errors where volume is highest.

Why this matters — real numbers and industry context

Automation in reconciliation isn’t theory — finance teams see large time savings when rules and matching get applied consistently. Industry studies and vendor analysis show automation can cut reconciliation or close-related manual work by large margins — typical estimates range from 30–80% time savings depending on the scope and maturity of the automation, and error rates decline materially as well. Use these as conservative planning figures when you build a case for enabling automation: start modest, measure results, and then expand.

A practical example that mirrors many real customers’ experiences: a mid-sized services firm processing ~300 monthly transactions reduced reconciliation time from roughly 20 hours a month to under five hours after implementing automated matching rules and scheduled runs — the finance team could then focus on exceptions and cash-flow analysis rather than line-by-line matching.

Practical implementation tips — from sandbox to production

Below are step-by-step suggestions (ordered by risk and impact) to put Odoo 19’s reconciliation improvements to work quickly.

1) Start in a sandbox and measure baseline

Clone a recent month’s bank statements into a test database and measure current manual time and exception counts. This baseline will make ROI visible and help tune matching rules.

2) Build a small set of automated reconciliation models first

Create 3–5 high-value models for the most frequent flows: customer payments that match invoices, supplier payments that match bills, and bank fees or small recurring write-offs. Use amount ranges, partner filters and label regex to make them precise. Enable Auto-validate only after you’ve verified matches in the sandbox.

3) Use regex and label matching sparingly and test thoroughly

Regex is powerful — use it for consistent invoice labels or bank narration patterns (e.g., INV-\d{5}) but validate on a broad sample to avoid false positives.

4) Leverage partner mapping and counterpart suggestions

Map frequent counterparty names and enable “suggest counterpart” when you want a human in the loop. For highly repetitive lines, move to full auto so they don’t appear at all.

5) Turn on scheduled automatic reconciliation for steady feeds

Set automatic reconciliation to run on bank imports and schedule a frequent job (hourly or daily depending on volume). This keeps the backlog small and avoids month-end spikes.

6) Reconcile draft entries carefully

If your business posts provisional entries, enable reconcile-draft for controlled flows. Remember that automatic moves created at reconcile time will be drafts — ensure your posting workflow handles that (approvals, validations).

7) Create an exception workflow and dashboards

Automation should flag only true exceptions. Build a simple queue (e.g., “Needs Review”) and a short checklist for approvers: check partner, verify dates, confirm amounts. Add an exceptions dashboard to see counts and trendlines.

8) Train, monitor, and iterate

When automation starts, review the first 2–4 weeks of auto-matches daily. Tweak model order (the sequence matters) and adjust regex/amount ranges. Over time the system’s suggestions will stabilize — keep a log of false positives to refine rules.

9) Audit trail and documentation for controls

Ensure every automatic match has an audit trail and a reason code. That makes internal audits painless and maintains control over who changed which rule and why.

10) Plan fallback and rollback

Keep a documented rollback procedure for each model (disable, reprocess the affected statement lines) so that a bad rule can be reversed quickly without manual reconstruction.

Common pitfalls and how to avoid them

- Too broad rules: Wide regex or empty partner filters cause false matches. Start narrow and broaden only when safe.

- Skipping human checks too soon: Don’t flip everything to auto-validate on day one. Use a phased approach.

- Ignoring posting workflow: If reconciliation creates draft moves, confirm your approvals and periodic posting rules handle them correctly.

Quick checklist to get started (first 30 days)

- Create a sandbox and import recent statements.

- Set up 3 core reconciliation models (invoices, bills, bank fees).

- Enable scheduled automatic reconciliation (low frequency to start).

- Review auto-matches daily and refine.

- Build an exceptions list and dashboard.

Final thought

Odoo 19’s reconciliation improvements are practical — they reduce repetitive work and let accountants do the judgement work that machines can’t. With a careful rollout (sandbox, narrow rules, scheduled runs, and a short review cycle), most teams see measurable time savings and steadier month-end closes.

Start small, measure results, and scale the rules that consistently reduce manual clicks and exceptions.