Sap Hana ML

What it does

When your agent uses it

• : Building ML solutions with SAP HANAs in-database machine learning using Python hana-ml library for PAL/APL algorithms
• DataFrame operations
• Model persistence

Example prompts

• “/sap-hana-ml”

Requirements

• Python 3

Workflow steps

6 steps, taken from the first numbered list in SKILL.md.

1. 1Use lazy evaluation - Operations build SQL without execution until collect()
2. 2Leverage in-database processing - Keep data in HANA for performance
3. 3Use Unified interfaces - Consistent APIs across algorithms
4. 4Save models - Use ModelStorage for persistence
5. 5Explain predictions - Use SHAP explainers for interpretability
6. 6Monitor AutoML - Use PipelineProgressStatusMonitor for long-running jobs

What it can do on your machine

Read from SKILL.md and the folder at commit 652a861. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Shell commands in SKILL.md call:

  • pip

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  Links to these hosts (documentation or services it may open):

  • help.sap.com
  • pypi.org

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Sap Hana ML loads about 1.7k tokens when it runs, and up to ~19k if it reads all its reference files. Until then it costs about 125 tokens; SKILL.md has 377 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~125

When it runs · the whole SKILL.md, loaded when a task matches

~1.7k

With references · SKILL.md plus every file in references/, read only if the agent opens them

~19k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from secondsky/sap-skills at commit 652a861, republished under its GPL-3.0 licence (© secondsky). 377 words, ~1,724 tokens.

name

sap-hana-ml

description

SAP HANA Machine Learning Python Client (hana-ml) development skill. Use when: Building ML solutions with SAP HANA's in-database machine learning using Python hana-ml library for PAL/APL algorithms, DataFrame operations, AutoML, model persistence, and visualization. Keywords: hana-ml, SAP HANA, machine learning, PAL, APL, predictive analytics, HANA DataFrame, ConnectionContext, classification, regression, clustering, time series, ARIMA, gradient boosting, AutoML, SHAP, model storage

license

GPL-3.0

metadata.maintainer

Eduard Jiglau

metadata.maintainer_email

hello@sap-ai-skills.com

metadata.website

https://sap-ai-skills.com

metadata.version

2.4.1

metadata.last_verified

2025-11-27

metadata.package_version

2.22.241011

pip install hana-ml

from hana_ml import ConnectionContext

# Connect
conn = ConnectionContext(
    address='<hostname>',
    port=443,
    user='<username>',
    password='<password>',
    encrypt=True
)

# Create DataFrame
df = conn.table('MY_TABLE', schema='MY_SCHEMA')
print(f"Shape: {df.shape}")
df.head(10).collect()

from hana_ml.algorithms.pal.unified_classification import UnifiedClassification

# Train model
clf = UnifiedClassification(func='RandomDecisionTree')
clf.fit(train_df, features=['F1', 'F2', 'F3'], label='TARGET')

# Predict & evaluate
predictions = clf.predict(test_df, features=['F1', 'F2', 'F3'])
score = clf.score(test_df, features=['F1', 'F2', 'F3'], label='TARGET')

from hana_ml.algorithms.apl.classification import AutoClassifier

# Automated classification
auto_clf = AutoClassifier()
auto_clf.fit(train_df, label='TARGET')
predictions = auto_clf.predict(test_df)

from hana_ml.model_storage import ModelStorage

ms = ModelStorage(conn)
clf.name = 'MY_CLASSIFIER'
ms.save_model(model=clf, if_exists='replace')

from hana_ml.algorithms.pal.partition import train_test_val_split

train, test, val = train_test_val_split(
    data=df,
    training_percentage=0.7,
    testing_percentage=0.2,
    validation_percentage=0.1
)

# APL models
importance = auto_clf.get_feature_importances()

# PAL models
from hana_ml.algorithms.pal.preprocessing import FeatureSelection
fs = FeatureSelection()
fs.fit(train_df, features=features, label='TARGET')

from hana_ml.algorithms.pal.pipeline import Pipeline
from hana_ml.algorithms.pal.preprocessing import Imputer, FeatureNormalizer

pipeline = Pipeline([
    ('imputer', Imputer(strategy='mean')),
    ('normalizer', FeatureNormalizer()),
    ('classifier', UnifiedClassification(func='RandomDecisionTree'))
])

from hana_ml.ml_exceptions import Error

try:
    clf.fit(train_df, features=features, label='TARGET')
except Error as e:
    print(f"HANA ML Error: {e}")

© secondsky, GPL-3.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 7 other files (references) in plugins/sap-hana-ml/skills/sap-hana-ml of secondsky/sap-skills.

• SKILL.md
• README.md
• agents/openai.yaml
• references/APL_ALGORITHMS.md
• references/DATAFRAME_REFERENCE.md
• references/PAL_ALGORITHMS.md
• references/SUPPORTING_MODULES.md
• references/VISUALIZERS.md

Open the folder on GitHubat commit 652a861