Emiliano

Estoy soltero

[Emiliano @Emilian40659768]

The tutorial covers the full pipeline: database design → bulk CSV import with auto-schema detection → SQL analytics (context by, pivot by) → asof/window joins for trading cost. All in #DolphinDB script—no external ETL, no Python round-trip. #DataEngineering #QuantFinance

[Emiliano @Emilian40659768]

pivot by turns long-form tick data into wide, time-aligned matrices in a single SQL clause. Group by minute, pivot by symbol, aggregate volume—and get a spreadsheet-ready table without leaving the database. No export. No reshape in Python. #DolphinDB #QuantFinance #SQL

[Emiliano @Emilian40659768]

context by is #DolphinDB's SQL extension that changes the game for panel data. Unlike group by, it returns the same number of rows as the input—so you can compute moving windows, cumulative metrics, and nested rolling correlations per symbol in one pass. #TimeSeries #SQL

[Emiliano @Emilian40659768]

Squared beta factor: rolling 60-month regression of asset excess returns on squared market excess returns. In CSAP it's one line—pow(mbeta(retrf, ewmktrf, 60, 20), 2)—using DolphinDB's built-in rolling regression. No hand-rolled loops. #DolphinDB #QuantFinance #RiskManagement

[Emiliano @Emilian40659768]

Vectorized parameter design across all 195 factors. Rolling-window and cross-sectional functions do the heavy lifting—mbeta, mavg, and friends—so factor code stays short, readable, and fast on time-series and panel data alike. #DolphinDB #QuantFinance #TimeSeries

[Emiliano @Emilian40659768]

"404 Project Not Found" → wrong repo ID. "401 Unauthorized" → bad access token. "No access to view deployModuleFromGitLab" → missing VIEW_EXEC grant. The tutorial maps 9 common errors to root causes—so you fix in minutes, not hours. #DolphinDB #Troubleshooting #DevOps

[Emiliano @Emilian40659768]

CSAP ships three modules: CSAPFactors (195 factors), CSAPPrepare (data cleaning + calc wrappers), CSAPDataSimulation (generate test data). No Compustat/CRSP access? Simulate with CSAPDataSimulation(gvkeyList, startYear, endYear) and start calculating. #DolphinDB #QuantFinance

[Emiliano @Emilian40659768]

GitLab + function views + modules = ordered, secure code management. Function views add user/group-level permissions. GitLab brings version control and collaboration. Modules keep directory-tree structure. One workflow, three problems solved. #DolphinDB #GitLab #CodeSecurity

[Emiliano @Emilian40659768]

Startup failure? Long-running transaction? Duplicate writes? DDBLogAnalyser ships purpose-built reports—startupAndShutdown, longestTransactions, transactionAudit—each isolating user, IP, and transaction ID to trace root cause fast. #DolphinDB #RootCause #Observability

[Emiliano @Emilian40659768]

Log investigations are tedious: manually hunting through files, guessing time ranges, cross-referencing nodes. #DolphinDB's DDBLogAnalyser module automates the whole loop—batch processing, precise time filtering, and report generation. #Observability #LogAnalysis

[Golden retriever @Goldennpaws]

Cute Golden retriever puppy ❤️😍

[Emiliano @Emilian40659768]

"Average return of the 10 highest-volume minutes in a 100-minute window"—previously a painful group-by filter. With mavgTopN(ratios(close), volume, 100, 10), it's one function call. Cleaner scripts, faster factor research. #DolphinDB #QuantFinance #AlphaMining

[Emiliano @Emilian40659768]

Built-in TopN not enough? aggrTopN(func, funcArgs, sortingCol, top) lets you plug in a custom aggregate—or wrap it in moving for sliding windows. Built-ins are faster, but the escape hatch exists when you need it. #DolphinDB #QuantFinance #AlphaMining

[Emiliano @Emilian40659768]

Three windowing flavors, one idea: mTopN for sliding windows by element count, tmTopN for time-based windows, cumTopN for cumulative history. All available across sum, avg, std, corr, beta, and more. #DolphinDB #TimeSeries #StreamProcessing

[Emiliano @Emilian40659768]

Direct vector calls? Roughly on par with Python TA-Lib (3ms vs 11ms for WMA on 1M elements). But grouped calculations are a different story: 17ms vs 535ms—31.5x faster. context by symbol is where the ta module pulls away. #DolphinDB #Performance #TimeSeries

[Emiliano @Emilian40659768]

#DolphinDB ships a full statistical toolkit: univariate measures (mad, skew, kurtosis, autocorr), multivariate covariance/correlation matrices, 10+ probability distributions, and random number generators—all built in, all ignoring nulls by default. #Statistics #TimeSeries

[Emiliano @Emilian40659768]

Rolling averages are where float errors accumulate quietly. With DECIMAL, mavg(decimal32(price, 4), 5) computes the exact moving average—no epsilon-tuning, no eqFloat workarounds. Same idea for cumulative returns via cumsum on DECIMAL64. #DolphinDB #QuantFinance #AlphaMining

[Doglover @puppiesDoglover]

Cute puppy

[Emiliano @Emilian40659768]

A #DolphinDB DECIMAL is stored as two parts: a raw integer and a scale. 1.23 at scale 2 is just the integer 123. Calculations run on the integers first, then apply the scale—so avg of [1.11, 2.22, 3.33] sums 111+222+333 before dividing. No drift. #DolphinDB #DataEngineering

[Emiliano @Emilian40659768]

Don't load into memory just to write it back. loadTextEx writes straight to the distributed database—384 ms vs 771 ms for the same file. The gap widens as data grows. #DolphinDB #BigData #Performance