DocumentatieExperience PlatformGids voor Query Service

Engineer-functies voor machinaal leren

Last update: Mon May 05 2025 00:00:00 GMT+0000 (Coordinated Universal Time)

Gemaakt voor:

  • Gebruiker
  • Ontwikkelaar

Dit document toont aan hoe u gegevens in Adobe Experience Platform in eigenschappen kunt omzetten, of variabelen, die door een machine het leren model kunnen worden verbruikt. Dit proces wordt bedoeld als eigenschapengineering. Met Data Distiller kunt u op schaal XML-functies berekenen en deze functies delen met uw computerleeromgeving. Dit betreft het volgende:

  1. Maak een querysjabloon om de doellabels en -functies te definiëren die u voor uw model wilt berekenen
  2. Voer de vraag uit en sla de resultaten in een trainingsdataset op

Uw trainingsgegevens definiëren define-training-data

Het volgende voorbeeld illustreert een vraag om opleidingsgegevens uit een dataset van de Gebeurtenissen van de Ervaring voor een model af te leiden om de neiging van een gebruiker te voorspellen om aan een nieuwsbrief in te tekenen. Abonnementsgebeurtenissen worden vertegenwoordigd door het gebeurtenistype web.formFilledOut en andere gedragsgebeurtenissen in de dataset worden gebruikt om functies op profielniveau af te leiden om abonnementen te voorspellen.

Positieve en negatieve labels voor query query-positive-and-negative-labels

Een volledige dataset voor de opleiding van een (onder toezicht staande) machinaal leermodel omvat doelvariabele of -label die het te voorspellen resultaat vertegenwoordigt, en een reeks eigenschappen of verklarende variabelen die worden gebruikt om de voorbeeldprofielen te beschrijven die worden gebruikt om het model op te leiden.

In dit geval is het label een variabele met de naam subscriptionOccurred die gelijk is aan 1 als het gebruikersprofiel een gebeurtenis van het type web.formFilledOut heeft en anders aan 0. De volgende vraag keert een reeks van 50.000 gebruikers van de gebeurtenissendataset, met inbegrip van alle gebruikers met positieve etiketten (subscriptionOccurred = 1) plus een reeks willekeurig geselecteerde gebruiker met negatieve etiketten terug om de 50.000 grootte van de gebruikerssteekproef te voltooien. Dit zorgt ervoor dat de trainingsgegevens zowel positieve als negatieve voorbeelden bevatten waarvan het model kan leren.

from aepp import queryservice

dd_conn = queryservice.QueryService().connection()
dd_cursor = queryservice.InteractiveQuery2(dd_conn)

query_labels = f"""
SELECT *
FROM (
    SELECT
        eventType,
        _{tenant_id}.user_id as userId,
        SUM(CASE WHEN eventType='web.formFilledOut' THEN 1 ELSE 0 END)
            OVER (PARTITION BY _{tenant_id}.user_id)
            AS "subscriptionOccurred",
        row_number() OVER (PARTITION BY _{tenant_id}.user_id ORDER BY randn()) AS random_row_number_for_user
    FROM {table_name}
)
WHERE (subscriptionOccurred = 1 AND eventType = 'web.formFilledOut') OR (subscriptionOccurred = 0 AND random_row_number_for_user = 1)
"""

df_labels = dd_cursor.query(query_labels, output="dataframe")
print(f"Number of classes: {len(df_labels)}")
df_labels.head()

de output van de Steekproef

Aantal klassen: 50000

eventType
userId
subscriptionOccurred
random_row_number_for_user
0
directMarketing.emailClicked
01027994177972439148069092698714414382
0
1
1
directMarketing.emailOpened
01054714817856066632264746967668888198
0
1
2
web.formFilledOut
01117296890525140996735553609305695042
1
15
3
directMarketing.emailClicked
01149554820363915324573708359099551093
0
1
4
directMarketing.emailClicked
01172121447143590196349410086995740317
0
1

Samengevoegde gebeurtenissen om functies voor ML te definiëren define-features

Met een aangewezen vraag kunt u de gebeurtenissen in de dataset in zinvolle, numerieke eigenschappen verzamelen die kunnen worden gebruikt om een aandrijvingsmodel te trainen. Voorbeelden van gebeurtenissen worden hieronder weergegeven:

  • Aantal e-mails die voor marketing doeleinden werden verzonden en door de gebruiker werden ontvangen.
  • Gedeelte van deze e-mails die ​ ​werden geopend.
  • Het gedeelte van deze e-mails waar de gebruiker ​ ​de verbinding selecteerde.
  • Aantal producten die werden bekeken.
  • Aantal voorstellingen die met werden in wisselwerking gestaan.
  • Aantal voorstellen die werden verworpen.
  • Aantal verbindingen die werden geselecteerd.
  • Aantal minuten tussen twee opeenvolgende ontvangen e-mails.
  • Aantal minuten tussen twee opeenvolgende e-mails geopend.
  • Het aantal minuten tussen twee opeenvolgende e-mailberichten waarin de gebruiker de koppeling heeft geselecteerd.
  • Aantal minuten tussen twee opeenvolgende productweergaven.
  • Het aantal minuten tussen twee voorstellingen waarmee interactie is opgetreden.
  • Aantal minuten tussen twee voorstellen die zijn afgewezen.
  • Aantal minuten tussen twee geselecteerde koppelingen.

De volgende query voegt deze gebeurtenissen samen:

Selecteren om voorbeeldquery weer te geven
code language-python 
query_features = f"""
SELECT
    _{tenant_id}.user_id as userId,
    SUM(CASE WHEN eventType='directMarketing.emailSent' THEN 1 ELSE 0 END)
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "emailsReceived",
    SUM(CASE WHEN eventType='directMarketing.emailOpened' THEN 1 ELSE 0 END)
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "emailsOpened",
    SUM(CASE WHEN eventType='directMarketing.emailClicked' THEN 1 ELSE 0 END)
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "emailsClicked",
    SUM(CASE WHEN eventType='commerce.productViews' THEN 1 ELSE 0 END)
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "productsViewed",
    SUM(CASE WHEN eventType='decisioning.propositionInteract' THEN 1 ELSE 0 END)
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "propositionInteracts",
    SUM(CASE WHEN eventType='decisioning.propositionDismiss' THEN 1 ELSE 0 END)
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "propositionDismissed",
    SUM(CASE WHEN eventType='web.webinteraction.linkClicks' THEN 1 ELSE 0 END)
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "webLinkClicks" ,
    TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'directMarketing.emailSent', 'minutes')
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "minutes_since_emailSent",
    TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'directMarketing.emailOpened', 'minutes')
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "minutes_since_emailOpened",
    TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'directMarketing.emailClicked', 'minutes')
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "minutes_since_emailClick",
    TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'commerce.productViews', 'minutes')
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "minutes_since_productView",
    TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'decisioning.propositionInteract', 'minutes')
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "minutes_since_propositionInteract",
    TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'propositionDismiss', 'minutes')
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "minutes_since_propositionDismiss",
    TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'web.webinteraction.linkClicks', 'minutes')
       OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
       AS "minutes_since_linkClick"
FROM {table_name}
"""

df_features = dd_cursor.query(query_features, output="dataframe")
df_features.head()

de output van de Steekproef

userId
emailsReceived
e-mailsGeopend
emailsClicked
productsViewed
propositionInteracts
voorstelAfgewezen
webLinkClicks
minutes_since_emailSent
minutes_since_emailOpened
minutes_since_emailClick
minutes_since_productView
minutes_since_propositionInteract
minutes_since_propositionDismiss
minutes_since_linkClick
0
01102546977582484968046916668339306826
1
0
0
0
0
0
0
0,0
NaN
NaN
NaN
NaN
Geen
NaN
1
01102546977582484968046916668339306826
2
0
0
0
0
0
0
0,0
NaN
NaN
NaN
NaN
Geen
NaN
2
01102546977582484968046916668339306826
3
0
0
0
0
0
0
0,0
NaN
NaN
NaN
NaN
Geen
NaN
3
01102546977582484968046916668339306826
3
1
0
0
0
0
0
540,0
0,0
NaN
NaN
NaN
Geen
NaN
4
01102546977582484968046916668339306826
3
2
0
0
0
0
0
588,0
0,0
NaN
NaN
NaN
Geen
NaN

Tags en functies combineren combine-queries

Tot slot kunnen de etiketvraag en de eigenschapvraag in één enkele vraag worden gecombineerd die een opleidingsdataset van etiketten en eigenschappen terugkeert:

Selecteren om voorbeeldquery weer te geven
code language-python 
query_training_set = f"""
SELECT *
FROM (
    SELECT _{tenant_id}.user_id as userId,
       eventType,
       timestamp,
       SUM(CASE WHEN eventType='web.formFilledOut' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id)
           AS "subscriptionOccurred",
       SUM(CASE WHEN eventType='directMarketing.emailSent' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "emailsReceived",
       SUM(CASE WHEN eventType='directMarketing.emailOpened' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "emailsOpened",
       SUM(CASE WHEN eventType='directMarketing.emailClicked' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "emailsClicked",
       SUM(CASE WHEN eventType='commerce.productViews' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "productsViewed",
       SUM(CASE WHEN eventType='decisioning.propositionInteract' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "propositionInteracts",
       SUM(CASE WHEN eventType='decisioning.propositionDismiss' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "propositionDismissed",
       SUM(CASE WHEN eventType='web.webinteraction.linkClicks' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "webLinkClicks" ,
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'directMarketing.emailSent', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_emailSent",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'directMarketing.emailOpened', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_emailOpened",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'directMarketing.emailClicked', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_emailClick",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'commerce.productViews', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_productView",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'decisioning.propositionInteract', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_propositionInteract",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'propositionDismiss', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_propositionDismiss",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'web.webinteraction.linkClicks', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_linkClick",
        row_number() OVER (PARTITION BY _{tenant_id}.user_id ORDER BY randn()) AS random_row_number_for_user
    FROM {table_name} LIMIT 1000
)
WHERE (subscriptionOccurred = 1 AND eventType = 'web.formFilledOut') OR (subscriptionOccurred = 0 AND random_row_number_for_user = 1)
ORDER BY timestamp
"""

df_training_set = dd_cursor.query(query_training_set, output="dataframe")
df_training_set.head()

de output van de Steekproef

userId
eventType
tijdstempel
subscriptionOccurred
emailsReceived
e-mailsGeopend
emailsClicked
productsViewed
propositionInteracts
voorstelAfgewezen
webLinkClicks
minutes_since_emailSent
minutes_since_emailOpened
minutes_since_emailClick
minutes_since_productView
minutes_since_propositionInteract
minutes_since_propositionDismiss
minutes_since_linkClick
random_row_number_for_user
0
02554909162592418347780983091131567290
directMarketing.emailSent
2023-06-17 13 :44: 59.086
0
2
0
0
0
0
0
0
0,0
NaN
NaN
NaN
NaN
Geen
NaN
1
1
01130334080340815140184601481559659945
directMarketing.emailOpened
2023-06-19 06 :01: 55.366
0
1
3
0
1
0
0
0
1921,0
0,0
NaN
1703,0
NaN
Geen
NaN
1
2
01708961660028351393477273586554010192
web.formFilledOut
2023-06-19 18 :36: 49.083
1
1
2
2
0
0
0
0
2365,0
26,0
1,0
NaN
NaN
Geen
NaN
7
3
01809182902320674899156240602124740853
directMarketing.emailSent
2023-06-21 19 :17: 12.535
0
1
0
0
0
0
0
0
0,0
NaN
NaN
NaN
NaN
Geen
NaN
1
4
03441761949943678951106193028739001197
directMarketing.emailSent
2023-06-21 21 :58: 29.482
0
1
0
0
0
0
0
0
0,0
NaN
NaN
NaN
NaN
Geen
NaN
1

Creeer een vraagmalplaatje om trainingsgegevens incrementeel te verwerken

Het is typisch om een model met bijgewerkte opleidingsgegevens periodiek om te herscholen om nauwkeurigheid van het model in tijd te handhaven. Als beste praktijken voor efficiënt het bijwerken van uw trainingsdataset, kunt u een malplaatje van uw vraag van de opleidingsreeks tot stand brengen om nieuwe opleidingsgegevens in oplopende mate te verwerken. Dit staat u toe gegevens verwerkt etiketten en eigenschappen slechts van gegevens die aan de originele dataset van de Gebeurtenissen van de Ervaring werden toegevoegd aangezien de opleidingsgegevens het laatst werden bijgewerkt, en neemt de nieuwe etiketten en de eigenschappen in de bestaande opleidingsdataset op.

Dit vereist een paar wijzigingen aan de vraag van de trainingsreeks:

  • Voeg logica toe om een nieuwe opleidingsdataset tot stand te brengen als het niet bestaat, en neem de nieuwe etiketten en de eigenschappen in de bestaande opleidingsdataset anders op. Hiervoor is een reeks van twee versies van de query voor de trainingsset vereist:

    • Eerst gebruikt u de instructie CREATE TABLE IF NOT EXISTS {table_name} AS
    • Vervolgens gebruikt u de instructie INSERT INTO {table_name} voor het geval waarin de trainingsdataset al bestaat

  • Voeg een instructie SNAPSHOT BETWEEN $from_snapshot_id AND $to_snapshot_id toe om de query te beperken tot gebeurtenisgegevens die binnen een opgegeven interval zijn toegevoegd. Het voorvoegsel $ op de momentopname-id's geeft aan dat het variabelen zijn die worden doorgegeven wanneer de querysjabloon wordt uitgevoerd.

Als u deze wijzigingen toepast, wordt de volgende query uitgevoerd:

Selecteren om voorbeeldquery weer te geven
code language-python 
ctas_table_name = "propensity_training_set"

query_training_set_template = f"""
$$ BEGIN

SET @my_table_exists = SELECT table_exists('{ctas_table_name}');

CREATE TABLE IF NOT EXISTS {ctas_table_name} AS
SELECT *
FROM (
    SELECT _{tenant_id}.user_id as userId,
       eventType,
       timestamp,
       SUM(CASE WHEN eventType='web.formFilledOut' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id)
           AS "subscriptionOccurred",
       SUM(CASE WHEN eventType='directMarketing.emailSent' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "emailsReceived",
       SUM(CASE WHEN eventType='directMarketing.emailOpened' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "emailsOpened",
       SUM(CASE WHEN eventType='directMarketing.emailClicked' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "emailsClicked",
       SUM(CASE WHEN eventType='commerce.productViews' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "productsViewed",
       SUM(CASE WHEN eventType='decisioning.propositionInteract' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "propositionInteracts",
       SUM(CASE WHEN eventType='decisioning.propositionDismiss' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "propositionDismissed",
       SUM(CASE WHEN eventType='web.webinteraction.linkClicks' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "webLinkClicks" ,
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'directMarketing.emailSent', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_emailSent",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'directMarketing.emailOpened', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_emailOpened",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'directMarketing.emailClicked', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_emailClick",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'commerce.productViews', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_productView",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'decisioning.propositionInteract', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_propositionInteract",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'propositionDismiss', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_propositionDismiss",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'web.webinteraction.linkClicks', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_linkClick",
        row_number() OVER (PARTITION BY _{tenant_id}.user_id ORDER BY randn()) AS random_row_number_for_user
    FROM {table_name}
    SNAPSHOT BETWEEN $from_snapshot_id AND $to_snapshot_id
)
WHERE (subscriptionOccurred = 1 AND eventType = 'web.formFilledOut') OR (subscriptionOccurred = 0 AND random_row_number_for_user = 1)
ORDER BY timestamp;

INSERT INTO {ctas_table_name}
SELECT *
FROM (
    SELECT _{tenant_id}.user_id as userId,
       eventType,
       timestamp,
       SUM(CASE WHEN eventType='web.formFilledOut' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id)
           AS "subscriptionOccurred",
       SUM(CASE WHEN eventType='directMarketing.emailSent' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "emailsReceived",
       SUM(CASE WHEN eventType='directMarketing.emailOpened' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "emailsOpened",
       SUM(CASE WHEN eventType='directMarketing.emailClicked' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "emailsClicked",
       SUM(CASE WHEN eventType='commerce.productViews' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "productsViewed",
       SUM(CASE WHEN eventType='decisioning.propositionInteract' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "propositionInteracts",
       SUM(CASE WHEN eventType='decisioning.propositionDismiss' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "propositionDismissed",
       SUM(CASE WHEN eventType='web.webinteraction.linkClicks' THEN 1 ELSE 0 END)
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "webLinkClicks" ,
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'directMarketing.emailSent', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_emailSent",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'directMarketing.emailOpened', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_emailOpened",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'directMarketing.emailClicked', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_emailClick",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'commerce.productViews', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_productView",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'decisioning.propositionInteract', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_propositionInteract",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'propositionDismiss', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_propositionDismiss",
       TIME_BETWEEN_PREVIOUS_MATCH(timestamp, eventType = 'web.webinteraction.linkClicks', 'minutes')
           OVER (PARTITION BY _{tenant_id}.user_id ORDER BY timestamp ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
           AS "minutes_since_linkClick",
        row_number() OVER (PARTITION BY _{tenant_id}.user_id ORDER BY randn()) AS random_row_number_for_user
    FROM {table_name}
    SNAPSHOT BETWEEN $from_snapshot_id AND $to_snapshot_id
)
WHERE
    @my_table_exists = 't' AND
    ((subscriptionOccurred = 1 AND eventType = 'web.formFilledOut') OR (subscriptionOccurred = 0 AND random_row_number_for_user = 1))
ORDER BY timestamp;

EXCEPTION
  WHEN OTHER THEN
    SELECT 'ERROR';

END $$;
"""

Tot slot bewaart de volgende code het vraagmalplaatje in Gegevens Distiller:

template_res = dd.createQueryTemplate({
  "sql": query_training_set_template,
  "queryParameters": {},
  "name": "Template for propensity training data"
})
template_id = template_res["id"]

print(f"Template for propensity training data created as ID {template_id}")

de output van de Steekproef

Template for propensity training data created as ID f3d1ec6b-40c2-4d13-93b6-734c1b3c7235

Met het bewaarde malplaatje, kunt u de vraag op elk ogenblik uitvoeren door malplaatjeidentiteitskaart van verwijzingen te voorzien en de waaier van momentopname IDs te specificeren die in de vraag zou moeten worden omvat. De volgende vraag wint de momentopnamen van de originele dataset van de Gebeurtenissen van de Ervaring terug:

query_snapshots = f"""
SELECT snapshot_id
FROM (
    SELECT history_meta('{table_name}')
)
WHERE is_current = true OR snapshot_generation = 0
ORDER BY snapshot_generation ASC
"""

df_snapshots = dd_cursor.query(query_snapshots, output="dataframe")

De volgende code toont uitvoering van het vraagmalplaatje aan, gebruikend eerste en laatste momentopnamen om de volledige dataset te vragen:

snapshot_start_id = str(df_snapshots["snapshot_id"].iloc[0])
snapshot_end_id = str(df_snapshots["snapshot_id"].iloc[1])

query_final_res = qs.postQueries(
    name=f"[CMLE][Week2] Query to generate training data created by {username}",
    templateId=template_id,
    queryParameters={
        "from_snapshot_id": snapshot_start_id,
        "to_snapshot_id": snapshot_end_id,
    },
    dbname=f"{cat_conn.sandbox}:all"
)
query_final_id = query_final_res["id"]
print(f"Query started successfully and got assigned ID {query_final_id} - it will take some time to execute")

de output van de Steekproef

Query started successfully and got assigned ID c6ea5009-1315-4839-b072-089ae01e74fd - it will take some time to execute

U kunt de volgende functie bepalen om de status van de vraag periodiek te controleren:

def wait_for_query_completion(query_id):
    while True:
        query_info = qs.getQuery(query_id)
        query_state = query_info["state"]
        if query_state in ["SUCCESS", "FAILED"]:
            break
        print("Query is still in progress, sleeping…")
        time.sleep(60)

    duration_secs = query_info["elapsedTime"] / 1000
    if query_state == "SUCCESS":
        print(f"Query completed successfully in {duration_secs} seconds")
    else:
        print(f"Query failed with the following errors:", file=sys.stderr)
        for error in query_info["errors"]:
            print(f"Error code {error['code']}: {error['message']}", file=sys.stderr)

wait_for_query_completion(query_final_id)

de output van de Steekproef

Query is still in progress, sleeping…
Query is still in progress, sleeping…
Query is still in progress, sleeping…
Query is still in progress, sleeping…
Query is still in progress, sleeping…
Query is still in progress, sleeping…
Query is still in progress, sleeping…
Query is still in progress, sleeping…
Query completed successfully in 473.8 seconds

Volgende stappen:

Door dit document te lezen hebt u geleerd hoe u gegevens in Adobe Experience Platform kunt transformeren in functies, of variabelen, die door een model voor machinaal leren kunnen worden verbruikt. De volgende stap in het creëren van eigenschappijpleidingen van Experience Platform om douanemodellen in uw machine het leren milieu te voeren is eigenschapdatasets van de uitvoer.

recommendation-more-help
ccf2b369-4031-483f-af63-a93b5ae5e3fb