Skip to content
IconMind

Data Engineering

Relational icons

50 icons in the relational group of Data Engineering: Auto increment, Cardinality, Cascade delete and 47 more — each in outline and duotone at three weights, with the code for React, Vue, Svelte, Flutter and more.

50 icons · 1 of 9 groupsOpen in browser

The 50 relational icons in Data Engineering, described

Auto incrementauto-increment
Auto-increment — the column that fills its own id, one higher each time.
Cardinalitycardinality
Cardinality — how many distinct values a column actually holds.
Cascade deletecascade-delete
Cascade delete — the children go when the parent does.
Check constraintcheck-constraint
A check constraint — a rule every row has to satisfy to be stored.
Collationcollation
Collation — the rule that decides how text sorts and compares.
Composite keycomposite-key
A composite key — a key made of more than one column together.
Connection poolconnection-pool
A connection pool — a handful of connections handed round instead of opened each time.
Cross joincross-join
A cross join — every row on the left paired with every row on the right.
CTEcte
A common table expression — a named result the main query then reads.
Default valuedefault-value
A default — the value written when the insert supplies none.
Distinctdistinct-rows
Distinct — duplicates collapsed so each row appears once.
Explain planexplain-plan
An explain plan — the tree of steps the planner chose to answer the query.
Full-text indexfull-text-index
A full-text index — words indexed so a phrase can be looked up, not just a value.
Group bygroup-by
Group by — rows gathered into buckets so each can be summarised.
Index bloatindex-bloat
Index bloat — an index grown far past the data it points at.
Covering indexindex-covering
A covering index — one that answers the query without touching the table.
Inner joininner-join
An inner join — only the rows that match on both sides.
Isolation levelisolation-level
Isolation level — how much of one transaction another is allowed to see.
Materialized viewmaterialized-view
A materialized view — a view whose answer is stored rather than recomputed.
N+1 queriesn-plus-one
The N+1 problem — one query for the list, then one more for every row in it.
Not nullnot-null-constraint
Not null — the column that refuses to be left empty.
Nullnull-value
Null — the absence of a value, which is not zero and not empty text.
Nullable columnnullable-column
A nullable column — one that is allowed to hold nothing at all.
Optimistic lockoptimistic-lock
Optimistic locking — write freely, then check the version still matches.
Order byorder-by
Order by — the rows handed back in the order the query asked for.
ORMorm
An ORM — objects on one side, tables on the other, mapped between.
Outer joinouter-join
An outer join — the matches, and the rows that had none, kept anyway.
Partitioned tablepartition-table
A partitioned table — one table stored as ranges that are queried as one.
Pessimistic lockpessimistic-lock
Pessimistic locking — take the lock first, and nobody else gets in.
Query cachequery-cache
A query cache — the answer kept so the same question is not asked twice.
Query hintquery-hint
A query hint — the directive that tells the planner which path to take.
Query timeoutquery-timeout
A query timeout — the statement cancelled because it ran too long.
Read replicaread-replica
A read replica — a copy kept in step that answers reads and takes no writes.
Referential integrityreferential-integrity
Referential integrity — every reference points at a row that exists.
Row lockrow-lock
A row lock — one row held while the rest of the table stays free.
Savepointsavepoint
A savepoint — a marked place inside a transaction to roll back to.
Seed dataseed-data
Seed data — the rows a fresh database is planted with.
Self joinself-join
A self join — a table joined to itself to compare its own rows.
Sequencesequence-db
A sequence — the generator that hands out the next number on request.
Slow queryslow-query
A slow query — the one over the threshold, found and timed.
Stored procedurestored-procedure
A stored procedure — logic that lives in the database and runs there.
Subquerysubquery
A subquery — a query nested inside the one that uses its answer.
Surrogate keysurrogate-key
A surrogate key — an id with no meaning of its own, standing in for one.
Table locktable-lock
A table lock — the whole table held, and everyone else waits.
Table scantable-scan
A table scan — every row read because no index could be used.
Transactiontransaction
A transaction — work that either all lands or none of it does.
Truncate tabletruncate-table
Truncate — every row swept out at once, the table itself left standing.
Unique constraintunique-constraint
A unique constraint — no two rows may carry the same value.
Vacuumvacuum-db
Vacuum — dead rows reclaimed and the space handed back.
Write-ahead logwrite-ahead-log
A write-ahead log — the change written to the log before it touches the page.

In code, each is one import — import { AutoIncrement } from "@iconmind/react/icons/auto-increment" — and the same name in Vue, Svelte, Solid, Preact, React Native, Astro, Blade and Flutter.

Narrow it by tag

Every relational icon in Data Engineering, free to ship

50 icons in outline and duotone at three weights, generated from one grid so nothing in the set can drift out of step. MIT licensed — commercial use, no attribution, no seat count.

Other groups in Data Engineering