Componenta Talend tMap: Tutorial Complet cu Exemple

// Mapare directa (drag and drop)
// row1.customer_id → output.customer_id
 
// Exemple de mapare bazata pe expresii:
 
// Transformari string
output.upper_name = StringHandling.UPCASE(row1.name)
output.lower_name = StringHandling.DOWNCASE(row1.name)
output.trimmed = StringHandling.TRIM(row1.name)
output.padded = StringHandling.LPAD(row1.code, 10, "0")
 
// Operatii substring
output.first_3 = StringHandling.LEFT(row1.code, 3)
output.last_4 = StringHandling.RIGHT(row1.phone, 4)
output.middle = StringHandling.MID(row1.code, 2, 4)
 
// Inlocuire string
output.cleaned = StringHandling.REPLACE(row1.text, "\n", " ")
output.no_spaces = StringHandling.EREPLACE(row1.text, "\\s+", "_")
 
// Concatenare
output.full_address = row1.street + ", " + row1.city + " " + row1.zip
 
// Gestionare null
output.safe_name = row1.name != null ? row1.name : "Unknown"
output.safe_amount = row1.amount == null ? 0 : row1.amount

// Matematica de baza
output.total = row1.quantity * row1.unit_price
output.discounted = row1.price * (1 - row1.discount_rate)
output.rounded = Math.round(row1.amount * 100.0) / 100.0
 
// Conversii de tip
output.int_value = Integer.parseInt(row1.string_number)
output.double_value = Double.parseDouble(row1.string_decimal)
output.string_value = String.valueOf(row1.numeric_field)
 
// Formatare numere
output.formatted = String.format("%.2f", row1.amount)
output.with_comma = String.format("%,.2f", row1.large_number)
output.percentage = String.format("%.1f%%", row1.rate * 100)
 
// Functii numerice
output.absolute = Math.abs(row1.value)
output.ceiling = Math.ceil(row1.decimal)
output.floor = Math.floor(row1.decimal)
output.max_value = Math.max(row1.val1, row1.val2)
output.min_value = Math.min(row1.val1, row1.val2)
output.power = Math.pow(row1.base, row1.exponent)
output.sqrt = Math.sqrt(row1.value)
 
// Valori aleatorii
output.random_int = (int)(Math.random() * 1000)
output.uuid = java.util.UUID.randomUUID().toString()

// Parseaza date din string-uri
output.parsed_date = TalendDate.parseDate("yyyy-MM-dd", row1.date_string)
output.parsed_datetime = TalendDate.parseDate("yyyy-MM-dd HH:mm:ss", row1.datetime_string)
 
// Formateaza date in string-uri
output.formatted_date = TalendDate.formatDate("MM/dd/yyyy", row1.date_field)
output.iso_date = TalendDate.formatDate("yyyy-MM-dd'T'HH:mm:ss'Z'", row1.datetime_field)
 
// Aritmetica pe date
output.next_week = TalendDate.addDate(row1.order_date, 7, "dd")
output.next_month = TalendDate.addDate(row1.start_date, 1, "MM")
output.next_year = TalendDate.addDate(row1.birth_date, 1, "yyyy")
output.yesterday = TalendDate.addDate(TalendDate.getCurrentDate(), -1, "dd")
 
// Extragerea partilor de data
output.year = TalendDate.getPartOfDate("YEAR", row1.date_field)
output.month = TalendDate.getPartOfDate("MONTH", row1.date_field)
output.day = TalendDate.getPartOfDate("DAY_OF_MONTH", row1.date_field)
output.day_of_week = TalendDate.getPartOfDate("DAY_OF_WEEK", row1.date_field)
output.week_of_year = TalendDate.getPartOfDate("WEEK_OF_YEAR", row1.date_field)
output.quarter = (TalendDate.getPartOfDate("MONTH", row1.date_field) - 1) / 3 + 1
 
// Comparatii de date
output.days_between = TalendDate.diffDate(row1.end_date, row1.start_date, "dd")
output.months_between = TalendDate.diffDate(row1.end_date, row1.start_date, "MM")
 
// Data/ora curenta
output.today = TalendDate.getCurrentDate()
output.now = TalendDate.parseDate("yyyy-MM-dd HH:mm:ss",
    TalendDate.formatDate("yyyy-MM-dd HH:mm:ss", new java.util.Date()))
 
// Prima/Ultima zi din luna
output.first_of_month = TalendDate.getFirstDayOfMonth(row1.date_field)
output.last_of_month = TalendDate.getLastDayOfMonth(row1.date_field)

/*
 * Configurare Lookup:
 *
 * Input Principal: row1 (customers)
 * Input Lookup: row2 (customer_segments)
 *
 * Cheie Join: row1.customer_id = row2.customer_id
 *
 * Lookup Model: Load once (pentru tabele lookup mici)
 * Tip Join: Left Outer Join
 * Model Potrivire: First Match
 */
 
// Expresii output cu date lookup
output.customer_id = row1.customer_id
output.customer_name = row1.name
output.segment = row2.segment != null ? row2.segment : "UNKNOWN"
output.tier = row2.tier != null ? row2.tier : "STANDARD"
 
// Gestioneaza lookup-uri fara rezultat cu valori implicite
output.discount_rate = row2 != null && row2.discount_rate != null
    ? row2.discount_rate
    : 0.0
 
// Conditional bazat pe rezultatul lookup
output.is_premium = row2 != null && "PREMIUM".equals(row2.tier)

/*
 * Configurare Lookup-uri Multiple:
 *
 * Input Principal: row1 (orders)
 * Lookup 1: row2 (customers) - join pe customer_id
 * Lookup 2: row3 (products) - join pe product_id
 * Lookup 3: row4 (regions) - join pe region_code
 *
 * Ordine procesare: Lookup-urile evaluate secvential
 */
 
// Sectiunea variabile (pentru reutilizare)
Var.customer_name = row2 != null ? row2.name : "Unknown Customer"
Var.product_name = row3 != null ? row3.product_name : "Unknown Product"
Var.region_name = row4 != null ? row4.region_name : "Unknown Region"
Var.unit_price = row3 != null ? row3.unit_price : row1.default_price
 
// Expresii output
output.order_id = row1.order_id
output.customer_name = Var.customer_name
output.product_name = Var.product_name
output.region = Var.region_name
output.line_total = row1.quantity * Var.unit_price
output.lookup_status = (row2 != null ? "C" : "") +
                       (row3 != null ? "P" : "") +
                       (row4 != null ? "R" : "")

/*
 * Lookup cu Cheie Compusa:
 *
 * Principal: row1 (transactions)
 * Lookup: row2 (price_matrix)
 *
 * Chei Join:
 *   row1.product_id = row2.product_id AND
 *   row1.region_code = row2.region_code AND
 *   row1.effective_date >= row2.start_date AND
 *   row1.effective_date <= row2.end_date
 *
 * Nota: Pentru intervale complexe de date, foloseste tMap cu expresie inner join
 * sau pre-filtreaza datele de lookup
 */
 
// Exemplu cheie expresie (in setarile cheie tMap)
// Key 1: row1.product_id.equals(row2.product_id)
// Key 2: row1.region_code.equals(row2.region_code)
 
// Pentru potrivire interval date, foloseste expresie:
row2.start_date.compareTo(row1.trans_date) <= 0 &&
row2.end_date.compareTo(row1.trans_date) >= 0

/*
 * Beneficii Sectiune Variabile:
 * - Evita repetarea expresiilor complexe
 * - Imbunatateste lizibilitatea
 * - Optimizeaza performanta (expresia evaluata o singura data)
 * - Construieste calcule pas cu pas
 */
 
// Definitii variabile (evaluate in ordine)
Var.base_price = row1.quantity * row3.unit_price
 
Var.discount_rate = row2 != null && row2.tier != null
    ? (row2.tier.equals("GOLD") ? 0.20 :
       row2.tier.equals("SILVER") ? 0.15 :
       row2.tier.equals("BRONZE") ? 0.10 : 0.05)
    : 0.0
 
Var.discount_amount = Var.base_price * Var.discount_rate
 
Var.subtotal = Var.base_price - Var.discount_amount
 
Var.tax_rate = row4 != null ? row4.tax_rate : 0.08
 
Var.tax_amount = Var.subtotal * Var.tax_rate
 
Var.total = Var.subtotal + Var.tax_amount
 
// Output folosind variabile
output.base_price = Var.base_price
output.discount_rate = Var.discount_rate
output.discount_amount = Var.discount_amount
output.subtotal = Var.subtotal
output.tax_rate = Var.tax_rate
output.tax_amount = Var.tax_amount
output.total = Var.total

// Asignare conditionala complexa
Var.customer_category =
    row1.annual_revenue > 1000000 ? "ENTERPRISE" :
    row1.annual_revenue > 100000 ? "MID_MARKET" :
    row1.annual_revenue > 10000 ? "SMALL_BUSINESS" :
    "MICRO"
 
Var.risk_level =
    row1.payment_days_late > 90 ? "HIGH" :
    row1.payment_days_late > 30 ? "MEDIUM" :
    row1.payment_days_late > 0 ? "LOW" :
    "NONE"
 
// Potrivire pattern cu regex
Var.is_valid_email = row1.email != null &&
    row1.email.matches("^[A-Za-z0-9+_.-]+@[A-Za-z0-9.-]+$")
 
// Logica state machine
Var.order_status =
    row1.shipped_date != null && row1.delivered_date != null ? "DELIVERED" :
    row1.shipped_date != null ? "IN_TRANSIT" :
    row1.paid_date != null ? "PROCESSING" :
    row1.confirmed_date != null ? "CONFIRMED" :
    "PENDING"

/*
 * Configurare Filtre in Setarile Output:
 *
 * Output 1 (all_orders): Fara filtru
 * Output 2 (large_orders): row1.amount > 1000
 * Output 3 (premium_customers): row2.tier.equals("PREMIUM")
 * Output 4 (needs_review): Var.needs_review == true
 */
 
// Expresii de filtru
 
// Filtru simplu pe valoare
row1.status.equals("ACTIVE")
 
// Comparatie numerica
row1.amount >= 1000 && row1.amount <= 10000
 
// Filtru pe data
row1.order_date.after(TalendDate.parseDate("yyyy-MM-dd", "2024-01-01"))
 
// Filtru verificare null
row1.email != null && !row1.email.isEmpty()
 
// Conditii multiple
row1.status.equals("ACTIVE") &&
row1.amount > 100 &&
row2 != null &&
row2.credit_limit > row1.amount
 
// Filtru pe pattern
row1.product_code.startsWith("PREM-")
 
// Apartenenta la lista
java.util.Arrays.asList("US", "CA", "MX").contains(row1.country_code)

/*
 * Configurare Flux Reject:
 *
 * Catch lookup reject: true
 * Catch expression filter reject: true
 *
 * Cauze reject:
 * - Lookup cu "Inner Join" si fara potrivire
 * - Expresia de filtru evaluata la false
 */
 
// In output-ul reject, poti accesa:
output.original_id = row1.id
output.original_data = row1.data
output.reject_reason = row2 == null ? "Customer lookup failed" :
    row3 == null ? "Product lookup failed" :
    "Filter condition not met"
output.rejected_at = TalendDate.getCurrentDate()

/*
 * Strategie de Rutare:
 *
 * Output Principal: Inregistrari reusite
 * Output Erori: Esecuri de validare
 * Output Arhiva: Inregistrari istorice
 * Output Delta: Doar inregistrari modificate
 */
 
// Folosirea variabilelor pentru decizii de rutare
Var.is_valid = row1.id != null &&
               row1.amount > 0 &&
               row1.customer_id != null
 
Var.is_historical = row1.record_date.before(
    TalendDate.addDate(TalendDate.getCurrentDate(), -365, "dd"))
 
Var.is_changed = !row1.hash.equals(row2.previous_hash)
 
// Filtre output:
// main_output: Var.is_valid && !Var.is_historical
// error_output: !Var.is_valid
// archive_output: Var.is_historical
// delta_output: Var.is_valid && Var.is_changed

// Defineste metoda reutilizabila in tJava (sectiunea begin)
public static String maskEmail(String email) {
    if (email == null || !email.contains("@")) return email;
    String[] parts = email.split("@");
    String username = parts[0];
    String masked = username.charAt(0) +
        "*".repeat(Math.max(0, username.length() - 2)) +
        (username.length() > 1 ? username.charAt(username.length() - 1) : "");
    return masked + "@" + parts[1];
}
 
public static String cleanPhone(String phone) {
    if (phone == null) return null;
    return phone.replaceAll("[^0-9]", "");
}
 
public static boolean isValidSSN(String ssn) {
    if (ssn == null) return false;
    return ssn.matches("^\\d{3}-\\d{2}-\\d{4}$");
}
 
// Foloseste in expresiile tMap:
output.masked_email = maskEmail(row1.email)
output.clean_phone = cleanPhone(row1.phone)
output.ssn_valid = isValidSSN(row1.ssn)

// GRESIT: Expresie complexa repetata
output.field1 = StringHandling.UPCASE(StringHandling.TRIM(row1.name))
output.field2 = StringHandling.UPCASE(StringHandling.TRIM(row1.name)) + "_SUFFIX"
output.field3 = StringHandling.UPCASE(StringHandling.TRIM(row1.name)).length()
 
// CORECT: Foloseste variabila
Var.clean_name = StringHandling.UPCASE(StringHandling.TRIM(row1.name))
output.field1 = Var.clean_name
output.field2 = Var.clean_name + "_SUFFIX"
output.field3 = Var.clean_name.length()
 
// GRESIT: Verificari null multiple
output.val = row2 != null ? (row2.field != null ? row2.field : "default") : "default"
 
// CORECT: Variabila singulara cu short-circuit
Var.lookup_value = row2 != null && row2.field != null ? row2.field : "default"
output.val = Var.lookup_value

/*
 * Selectie Lookup Model:
 *
 * 1. Load Once
 *    - Intregul lookup incarcat in memorie la inceput
 *    - Cel mai bun pentru: < 100K randuri, accesat frecvent
 *    - Memorie: O(n) unde n = dimensiunea lookup
 *
 * 2. Reload at Each Row
 *    - Interogeaza lookup-ul pentru fiecare rand principal
 *    - Cel mai bun pentru: Lookup-uri foarte mari, flux principal selectiv
 *    - Performanta: Lent, dar memorie scazuta
 *
 * 3. Reload at Each Row (cache)
 *    - Interogeaza + cache rezultate
 *    - Cel mai bun pentru: Lookup mare, chei repetate in principal
 *    - Memorie: O(chei unice accesate)
 */
 
// Pentru lookup-uri uriase, considera:
// 1. Pre-filtreaza datele lookup cu tFilterRow
// 2. Foloseste componente tHash* pentru lookup-uri bazate pe disc
// 3. Imparte job-ul in chunk-uri cu tFlowToIterate

tmap_best_practices:
  design:
    - "Foloseste nume semnificative pentru variabile"
    - "Comenteaza expresiile complexe"
    - "Pastreaza expresiile simple si lizibile"
    - "Imparte logica complexa in variabile multiple"
 
  performanta:
    - "Alege modelul de lookup corespunzator pe baza dimensiunii datelor"
    - "Foloseste variabile pentru expresii repetate"
    - "Filtreaza devreme pentru a reduce procesarea"
    - "Pre-filtreaza lookup-urile mari inainte de tMap"
 
  gestionare_erori:
    - "Gestioneaza intotdeauna valorile null explicit"
    - "Foloseste output-uri reject pentru tracking erori"
    - "Valideaza tipurile de date inainte de conversie"
    - "Logheaza inregistrarile rejectate pentru analiza"
 
  mentenanta:
    - "Documenteaza cheile de join si logica de business"
    - "Testeaza cu cazuri limita (null-uri, string-uri goale)"
    - "Versioneaza configuratiile tMap"
    - "Revizuieste si optimizeaza periodic"