1.7.0

✨ What's Changed

🚀 feat: In dh.getRmStats now return summary of missingness too by @timcadman in #143

🌐 Full Changelog: 1.6.0...1.7.0

v1.6.0

✨ What's Changed

• 🔧 Fix: Removed hard coding in dh.getRmStats by @timcadman in #141
• 🚨 Deprecation: Following release of dsTidyverse many dsHelper functions have been deprecated by @timcadman in #142:
  • dh.columnCast
  • dh.defineCases
  • dh.dropCols
  • dh.makeAgePolys (superceded)
  • dh.makeStrata
  • dh.renameVars
  • dh.tidyEnv

🌐 Full Changelog: Compare 1.5.1...1.6.0

1.5.1

Bugs

• Fixed error in calculation of pooled_se in ds.pool: previously was returning variance and not standard error

Full Changelog: 1.5.0...1.5.1

1.5.0

What's Changed

Features

• Added function dh.pool to pool estimates from across multiply imputed datasets using Rubin's Rules.
• Add functionality to test for single polynomial models in dh.makeLmerFormula.
• In dh.createTableOne Increased default value for significant digits to 3.
• In dh.createTableOne removed default for continuous variable format so the researcher has to explicitly choose.
• In dh.metaSepModels now return additional statistics.
• In dh.getStats, can now leave vars as NULL to return stats for all variables in a data frame.

Bugs

• Fixed some bugs in dh.createTableOne
• In ds.tidyEnv update code to no longer return a warning with dsBaseClient v6.3

Full Changelog: 1.1.0...1.5.0

1.1.0

Features

• Added dh.buildModels to build regression formula as part of a pipeline with dh.anyData and dh.multGLM.
• Added dh.getRmStats as an analogue of dh.getStats, which returns stats for repeated measures data (e.g. number of observations per participant).
• Added dh.stablisedWeights to calculate stablised IPW weights.
• In dh.multGlm now added argument weights to allow user to specify serverside object containing regression weights.

Fix

• In dh.anyData, stopped error being thrown if variable didn't exist within data frame. Instead reports that there is not data on that variable.
• In dh.lmTab changed argument name from exp to exponentiate. This was to fix a bug caused by conflict with R base function exp.
• In dh.lmTab corrected input validation to return correct error messages.
• In dh.lmTab fixed bug where it failed for type = "binomial" due to name of 'p-value' column.
• In dh.makeLmerForm fixed bug whereby formula was not created correctly if arguments fixed and age_interactions were both NULL.
• In dh.predictLmer fixed bug whereby due to sorting of lists the SEs and predictions were not joined correctly for each cohort.
• In dh.makeStrata modified utils function .checkDisclosureso that it works correctly with both dsBaseClient v.6.3 and below. Changes made to ds.table had broken this function.

Docs

• Changed name from dt.buildModels to dh.buildModels to match naming convention of package.
• Updated Trajectories vignette and pre-rendered correctly.
• Fixed other vignettes which were also not correctly pre-rendered.

Style

• Styled package: this had not been done for a while because the styler package is not currently working on the Central Analysis Server.

Refactor

• Completely refactored dh.predictLmer.

Chores

• In dh.tidyEnv, fixed the name of the argument x.names in the function ds.rm called within.
• Removed non-standard folder 'figures' and moved all figures created through pre-rendering vignettes into the 'vignettes' folder.

Test

• In dh.lmTab and dh.makeLmerForm fixed unit tests.

1.0.3

Fix

• In dh.makeIqr, added 'stringsAsFactors argument to call to ds.dataFrame to be compatible with latest version of R.
• In dh.makeExcludedDf, removed hard coding of ID.
• In dh.lmeMultPoly, fixed error where fit.tab wasn't being returned if there was only one valid model.

Chore

• Refactored dh.lmTab and dh.makeLmerForm.
• Added unit tests for dh.lmTab and dh.makeLmerForm.

1.0.2

Unintended increment - ignore this release

1.0.1

Unintended increment - ignore this release

1.0.0

Features

• In dh.metaSepModels increased the default number of digits returned so sufficient information is retained.
• In dh.getStats now return empty tibble rather than empty list if no continuous or categorical variables are provided.

## Bug fixes

• In dh.makeAgePolys added missing conns argument within function.
• In dh.makeLmerForm set default argument of age_interactions` to NULL.
• In dh.lmTab fixed a couple of bugs, where 'cohort' info was not returned where ci_format = paste, and also was breaking for option 'lmer' when there was no pvalue.
• In dh.dropCols ensured that argument new_obj has a default of null so that it doesn't take value from environment.
• In dh.createTableOne completely refactored and fixed a number of errors.
• In dh.getStats fixed error with checking for same levels across input variables.
• In dh.dropCols, dh.makeStrata, dh.renameVars fixed bug created by upgraded version of R in some cohorts. StringsAsFactors argument now set to FALSE for all calls to ds.dataFrame.

Performance

• In dh.lmeMultPoly refactored slightly to improve performance.
• In dh.predictLmer refactored slightly to improve performance.

Defunct

• dh.meanByAge is now defunct: use dh.meanByGroup instead.
• dh.makeOutcome is now defunct: use dh.makeStrata instead.

0.4.21

Features

• Added new function dh.createTableOne to create tables of descriptive statistics that can be used in manuscripts.
• Added new function dh.makeExcludedDf to create df of excluded participants, useful in describing missingness.
• Added new function dh.metaSepModels which meta-analyses models performed separately by cohort.
• Added new function dh.multGlm which loops over ds.glmSLMA and saves error output in helpful way.
• Added new function dh.zScoreByGroup which calculates z scores in repeated measures data by a grouping variable.
• In dh.lmTab now also return pvalue for slma option
• in dh.lmeMultPoly now label the returned models
• in dh.makeAgePolys now include missing conns argument for cbind function call
• in dh.meanByGroup now return additional descriptives
• In dh.predictLmer now add option whether or not to include intercept in newdata

Bug fixes

• in dh.castCols now imports 'tail'
• in dh.defineCases fixes typo in ifelse statements
• in dh.dropCols now defaults new_obj to original df
• in dh.getStats, fix an error whereby the returned cohort n was incorrect
• in dh.makeLmerForm fixed error in argument default

Performance

• in dh.makeAgePolys now remove temporary objects