MakeTimestamp logo
MakeTimestamp

Tutorial

Como converter carimbos de data/hora em SQL

Introdução

Trabalhar com carimbos de data/hora em bancos de dados SQL é uma habilidade fundamental para desenvolvedores e DBAs. Este guia aborda funções de conversão nas principais plataformas de banco de dados, considerações de desempenho e armadilhas comuns.

Referência rápida

Funções de conversão por banco de dados

MySQL

-- Convert Unix timestamp to DATETIME
SELECT
  id,
  UNIX_TIMESTAMP(created_at) AS converted_date
FROM events;

-- Convert DATETIME to Unix timestamp
SELECT
  id,
  UNIX_TIMESTAMP(created_at) AS timestamp_value
FROM events;

-- Current timestamp in Unix format
SELECT UNIX_TIMESTAMP(NOW()) AS current_unix;

Importante: MySQL TIMESTAMP tem um problema de 2038 (intervalo: 1970-01-01 00:00:00 a 2038-01-19 03:14:07). Use DATETIME ou BIGINT para datas além desse intervalo.

###PostgreSQL

-- Convert Unix timestamp to TIMESTAMP WITH TIME ZONE
SELECT
  id,
  to_timestamp(event_timestamp) AS converted_date
FROM events;

-- Convert timestamp to TIMESTAMPTZ with timezone
SELECT
  id,
  event_timestamp AT TIME ZONE 'America/New_York' AS eastern_time
FROM events;

-- Current timestamp in Unix format
SELECT EXTRACT(EPOCH FROM NOW()) AS current_unix;

Dica: O PostgreSQL TIMESTAMP COM TIME ZONE reconhece o fuso horário. Use AT TIME ZONE para conversões com suporte de fuso horário.

SQL Server (T-SQL)

-- Convert Unix timestamp to DATETIME2 (higher precision)
SELECT
  id,
  DATEADD(s, 1970-01-01 00:00:00, UNIX_TIMESTAMP(timestamp_column)) AS converted_date
FROM events;

-- Convert DATETIME2 to Unix timestamp
SELECT
  id,
  DATEDIFF(s, 1970-01-01 00:00:00, GETDATE()) AS unix_seconds
FROM events;

-- Current timestamp in Unix format
SELECT DATEDIFF(s, 1970-01-01 00:00:00, GETUTCDATE()) AS current_unix;

Observação: DATETIME2 fornece segundos fracionários (3 casas decimais) para maior precisão. Use a combinação DATEDIFF/DATEADD para cálculos precisos.

###SQLite

-- Convert Unix timestamp to DATETIME
SELECT
  id,
  datetime(timestamp_column, 'unixepoch') AS converted_date
FROM events;

-- Current timestamp in Unix format
SELECT strftime('%s', 'now') AS current_unix;

Nota: SQLite usa época Unix por padrão. Use modificadores strftime para formatos personalizados.

Oráculo

-- Convert Unix timestamp to DATE
SELECT
  id,
  TO_DATE('1970-01-01', 'YYYY-MM-DD') AS converted_date
FROM events;

-- Current timestamp
SELECT TO_CHAR(SYSTIMESTAMP, 'YYYY-MM-DD HH24:MI:SS') AS current_timestamp
FROM DUAL;

Importante: O formato TO_DATE do Oracle diferencia maiúsculas de minúsculas. Sempre use especificadores de formato em maiúsculas (AAAA, MM, DD, etc.).

Exemplos práticos

Exemplo 1: Converter e exibir

-- MySQL: Convert and format in single query
SELECT
  id,
  event_name,
  DATE_FORMAT(UNIX_TIMESTAMP(created_at), '%Y-%m-%d %H:%i') AS formatted_date,
  UNIX_TIMESTAMP(created_at) AS unix_value
FROM events
WHERE created_at >= DATE_SUB(NOW(), INTERVAL 7 DAY)
ORDER BY created_at DESC;

Exemplo 2: Filtrar por intervalo de datas

-- PostgreSQL: Efficient date range filtering with timestamp
SELECT
  id,
  event_name,
  event_timestamp AT TIME ZONE 'UTC' AS utc_time,
  TO_CHAR(event_timestamp AT TIME ZONE 'UTC', 'YYYY-MM-DD HH24:MI') AS formatted_date
FROM events
WHERE event_timestamp >= EXTRACT(EPOCH FROM (NOW() - INTERVAL '7 days'))
  AND event_timestamp < EXTRACT(EPOCH FROM NOW())
ORDER BY event_timestamp DESC;

Exemplo 3: Conversão em lote com vários formatos

-- SQL Server: Provide multiple timestamp formats in one query
SELECT
  id,
  event_name,
  event_timestamp,
  CONVERT(VARCHAR, event_timestamp, 120) AS iso8601,  -- Truncate to 120 chars
  CONVERT(DATETIME2, event_timestamp) AS datetime2,
  CONVERT(DATETIME, event_timestamp) AS readable_date,
  YEAR(event_timestamp) AS year,
  MONTH(event_timestamp) AS month,
  DAY(event_timestamp) AS day
FROM events
WHERE event_timestamp IS NOT NULL;

Exemplo 4: Cálculo da diferença horária

-- MySQL: Calculate time difference in seconds
SELECT
  id1, id2,
  timestamp1, timestamp2,
  TIMESTAMPDIFF(SECOND, timestamp2, timestamp1) AS diff_seconds,
  TIMESTAMPDIFF(DAY, timestamp2, timestamp1) AS diff_days,
  SEC_TO_TIME(TIMESTAMPDIFF(SECOND, timestamp2, timestamp1)) AS time_diff
FROM events
WHERE id IN (1, 2);

Otimização de desempenho

Estratégias de Índice

Índice primário na coluna Timestamp

Prática recomendada: sempre crie um índice primário em colunas de carimbo de data/hora para consultas de intervalo e operações de classificação.

-- MySQL: Primary index for range queries
CREATE TABLE events (
  id INT PRIMARY KEY,
  event_name VARCHAR(255),
  event_timestamp TIMESTAMP,
  INDEX idx_timestamp (event_timestamp)
);

-- PostgreSQL: Partial index for range queries
CREATE TABLE events (
  id BIGSERIAL PRIMARY KEY,
  event_name VARCHAR(255),
  event_timestamp TIMESTAMPTZ,
  INDEX idx_timestamp_range (event_timestamp)
);

-- SQL Server: Include timestamp in covering index
CREATE NONCLUSTERED INDEX idx_event_timestamp
ON events (event_timestamp DESC);

Índices Funcionais

Prática recomendada: Crie índices funcionais em colunas computadas (ano, mês, dia) para consultas de agrupamento de datas.

-- MySQL: Functional index for date-based grouping
CREATE TABLE events (
  id INT PRIMARY KEY,
  event_name VARCHAR(255),
  event_timestamp TIMESTAMP,
  INDEX idx_year (YEAR(event_timestamp)),
  INDEX idx_month (MONTH(event_timestamp)),
  INDEX idx_day (DAY(event_timestamp))
);

-- PostgreSQL: Generated column for automatic year/month/day
CREATE TABLE events (
  id BIGSERIAL PRIMARY KEY,
  event_name VARCHAR(255),
  event_timestamp TIMESTAMPTZ,
  event_year INTEGER GENERATED ALWAYS AS (EXTRACT(YEAR FROM event_timestamp)) STORED,
  event_month INTEGER GENERATED ALWAYS AS (EXTRACT(MONTH FROM event_timestamp)) STORED
);

Otimização de consulta

Evite chamadas de função em colunas

Erro comum: Usar funções como YEAR(), MONTH(), DAY() em colunas indexadas evita o uso do índice.

-- BAD: Function call prevents index usage
SELECT id, event_name
FROM events
WHERE YEAR(event_timestamp) = 2024
AND MONTH(event_timestamp) = 6;

-- GOOD: Compare literal values to use index
SELECT id, event_name
FROM events
WHERE event_timestamp >= '2024-01-01 00:00:00'
  AND event_timestamp < '2024-07-01 00:00:00';

Use parâmetros compatíveis com SARGABLE

Dica do MySQL: A partir do MySQL 8.0.18, o uso de parâmetros SARGABLE em instruções preparadas permite o uso do índice.

-- MySQL: Create prepared statement for parameterized queries
PREPARE stmt FROM 'SELECT * FROM events WHERE event_timestamp >= ? AND event_timestamp < ?';

-- Execute with parameters (efficient)
EXECUTE stmt USING @start_ts, @end_ts;

Limitar conjuntos de resultados

Prática recomendada: Sempre use LIMIT com consultas de intervalo para retornar apenas os dados necessários.

-- MySQL: Paginate large datasets
SELECT id, event_name, event_timestamp
FROM events
WHERE event_timestamp >= UNIX_TIMESTAMP('2025-01-01')
ORDER BY event_timestamp ASC
LIMIT 1000;

-- PostgreSQL: Use cursor-based pagination for large result sets
DECLARE cursor CURSOR FOR
  SELECT id, event_name, event_timestamp
  FROM events
  WHERE event_timestamp >= '2025-01-01 00:00:00 UTC'
  ORDER BY event_timestamp ASC;

Armadilhas Comuns

Tratamento de fuso horário

Funções de fuso horário do MySQL

Importante: MySQL TIMESTAMP NÃO armazena informações de fuso horário. Use TIMESTAMP WITH TIME ZONE para armazenamento com reconhecimento de fuso horário ou use DATETIME com coluna de fuso horário separada.

-- Convert to specific timezone
SELECT
  id,
  event_name,
  event_timestamp AT TIME ZONE 'America/Los_Angeles' AS la_time,
  event_timestamp AT TIME ZONE 'America/New_York' AS ny_time
FROM events;

-- Get current time in specific timezone
SELECT NOW() AS la_current, CONVERT_TZ(UTC, 'America/New_York') AS ny_current;

Alternativa: Armazene o carimbo de data/hora UTC e uma coluna de fuso horário separada para aplicativos com vários fusos horários.

Integridade de dados

Garanta a consistência do carimbo de data/hora

-- MySQL: Add CHECK constraint for reasonable timestamp range
CREATE TABLE events (
  id INT PRIMARY KEY,
  event_timestamp TIMESTAMP NOT NULL,
  CHECK (event_timestamp BETWEEN
    UNIX_TIMESTAMP('1970-01-01') AND
    UNIX_TIMESTAMP('2038-01-19')
  )
);

-- PostgreSQL: Use EXCLUDE constraint to prevent invalid data
CREATE TABLE events (
  id BIGSERIAL PRIMARY KEY,
  event_timestamp TIMESTAMPTZ NOT NULL,
  EXCLUDE (
    event_timestamp < EXTRACT(EPOCH FROM TIMESTAMP '1970-01-01')
    OR event_timestamp > EXTRACT(EPOCH FROM TIMESTAMP '2038-01-19')
  )
);

Estratégias de migração

Migração Incremental

Estratégia: Para tabelas grandes, migre em lotes usando cláusulas WHERE baseadas em carimbo de data/hora para evitar transações de longa duração.

-- Process 1000 rows at a time, ordered by timestamp
UPDATE events
SET last_processed = 1
WHERE event_timestamp < (
  SELECT event_timestamp
  FROM events
  WHERE last_processed = 0
  ORDER BY event_timestamp ASC
  LIMIT 1000
);

-- Continue until all rows processed
-- Repeat until all events marked as processed

Padrões Avançados

Tratamento de carimbos de data e hora da época

Padrões específicos do banco de dados

Padrões MySQL

Unix Timestamp com precisão de microssegundos

-- MySQL: Using BIGINT for millisecond timestamps
CREATE TABLE high_precision_events (
  id BIGINT PRIMARY KEY,
  event_timestamp BIGINT,  -- Milliseconds since epoch
  event_micros INT(3),  -- Microseconds (0-999)
  INDEX idx_timestamp (event_timestamp DESC)
);

-- Convert millisecond timestamp to human-readable format
SELECT
  id,
  FROM_UNIXTIME(event_timestamp / 1000) AS seconds,
  DATE_FORMAT(FROM_UNIXTIME(event_timestamp / 1000), '%Y-%m-%d %H:%i:%s') AS formatted
FROM high_precision_events;

Padrões PostgreSQL

Usando EXTRACT para componentes de data

-- PostgreSQL: Extract date components for filtering
SELECT
  id,
  event_name,
  EXTRACT(YEAR FROM event_timestamp) AS year,
  EXTRACT(MONTH FROM event_timestamp) AS month,
  EXTRACT(DAY FROM event_timestamp) AS day,
  DATE_TRUNC('day', event_timestamp) AS date_only
FROM events
WHERE event_name = 'Daily Backup';

Resumo das melhores práticas

Lista de verificação de desempenho

✅ Use appropriate data types (TIMESTAMP vs DATETIME)
✅ Create indexes on timestamp columns
✅ Use SARGABLE-able parameters when possible
✅ Limit result sets with pagination
✅ Avoid function calls on indexed columns
✅ Use WHERE clauses on timestamp for range queries
✅ Consider generated columns for date filtering
✅ Validate timestamp ranges on input
✅ Handle timezones explicitly (don't rely on implicit conversion)
✅ Use CHECK/EXCLUDE constraints for data integrity
❌ Don't store both timestamp and datetime for same data
❌ Don't use VARCHAR for timestamp columns
❌ Don't convert timestamp to string for every query

Para aplicativos de carimbo de data/hora de alto volume (>1 milhão de linhas), a indexação adequada e os padrões de consulta podem reduzir o tempo de consulta em 50-80%.

Ferramentas relacionadas