flextable 0.7.2 is out

David Gohel - 2022/06/23

tabulator
- Summarizor for table summary
Automating flextable printing
- data.frames as flextables
- Models as flextables
Separate column names into multiple rows
GitHub Discussions

The latest version of ‘flextable’ is on CRAN since few days.

This version comes with some major improvements:

soft returns \n are now supported in autofit(),
the HTML rendering of vertically oriented text is now correct and consistent with horizontal alignments,
tabs \t and soft returns \n are now correctly handled in PowerPoint,
it is now possible to place a same reference to a footnote several times.
the PDF rendering allows three placement options, fixed, float and wrap.

It also contains some new features that we are happy to introduce in this post.

tabulator

This new function prepares aggregations in a layout that can be turned into a nice flextable. The objective is to provide a tool to allow the creation of reporting tables where dimensions are arranged in rows and columns and where one or more statistics are displayed in the cell contents. ‘flextable’ grammar can be used to customize the table generated with comamnd flextable::as_flextable().

The function will create this table from already calculated aggregations.

Table 1: Prices of over 50 000 round cut diamonds
Subset of original dataset
Good n = 2634	D		3.5 (± 2.2)		3.0 (± 2.8)		3.6 (± 3.4)
E		4.4 (± 2.5)		3.2 (± 3.1)		3.8 (± 3.3)
F		2.6 (± 1.9)		3.3 (± 3.0)		3.8 (± 3.2)
G		3.2 (± 2.0)		4.1 (± 3.9)		4.1 (± 3.3)
H		3.8 (± 2.2)		4.2 (± 3.8)		4.4 (± 4.2)
I		4.2 (± 2.9)		4.7 (± 4.5)		6.0 (± 5.1)
J		3.8 (± 2.1)		4.6 (± 3.9)		4.8 (± 3.6)
Very Good n = 5915	D		2.6 (± 0.8)		3.2 (± 2.9)		3.1 (± 3.3)
E		3.4 (± 2.2)		3.2 (± 3.1)		3.3 (± 3.5)
F		4.3 (± 2.6)		3.6 (± 3.2)		4.0 (± 3.8)
G		3.2 (± 2.1)		3.5 (± 3.6)		4.4 (± 4.0)
H		5.3 (± 2.5)		4.9 (± 4.4)		4.6 (± 4.0)
I		6.0 (± 4.9)		5.2 (± 4.5)		5.8 (± 5.1)
J		4.5 (± 2.7)		5.0 (± 4.1)		5.3 (± 4.5)

In a Word document, the result would be the following. Note the page number is indicated in the top row of the table. The code for this example is available in the flextable gallery here.

‘tabulator’ benefits for all flextable grammar and can build complex tables made with graphics and text:

Summarizor for table summary

Function summarizor() is an utility function that prepare a numerical summary for each variables of a dataset, eventually by groups. It is supposed to be used as argument of tabulator().

The code for this example is available in the flextable gallery here.

Automating flextable printing

flextable provides two new functions dedicated to ‘R Markdown’ that turn automatically data.frame and models to flextable.

data.frames as flextables

We brought a code that we were using for our trainings that allows to display data.frame as flextables inside R Markdown documents. We tried to mimic the output of ‘data.table’ and ‘tibble’ by displaying just the first few rows and the type of the columns. To take advantage of this feature, we need to call the use_df_printer() function and voila!

use_df_printer()
airquality

Ozone	Solar.R	Wind	Temp	Month	Day
41	190	7.4	67	5	1
36	118	8.0	72	5	2
12	149	12.6	74	5	3
18	313	11.5	62	5	4
		14.3	56	5	5
28		14.9	66	5	6
23	299	8.6	65	5	7
19	99	13.8	59	5	8
8	19	20.1	61	5	9
	194	8.6	69	5	10
n: 153

Models as flextables

We have implemented the reporting of models into flextable, the function as_flextable() is doing the transformation. For now, the following models are supported: gam, lm, glm, glmmTMB, glmmadmb, htest, kmeans, pam, lme, gls, merMod, nlme and brmsfit.

This is an example made with a lme model:

library(nlme)
m1 <- lme(distance ~ age, data = Orthodont)
ft <- as_flextable(m1)
ft

group		Estimate	Standard Error	df	statistic	p-value
Fixed effects
	(Intercept)	16.761	0.775	80	21.620	0.0000	***
	age	0.660	0.071	80	9.265	0.0000	***
Random effects
Subject	sd_(Intercept)	2.327
Subject	cor_age.(Intercept)	-0.609
Subject	sd_age	0.226
Residual	sd_Observation	1.310
Signif. codes: 0 <= '*' < 0.001 < '' < 0.01 < '*' < 0.05

square root of the estimated residual variance: 1.3
data's log-likelihood under the model: -221.3
Akaike Information Criterion: 454.6
Bayesian Information Criterion: 470.6

In a R Markdown document, it is possible to automate the printing of models as flextable by using the function use_model_printer(). This is an example made with a kmean clustering:

use_model_printer()
kmeans(scale(mtcars[1:7]), 5)

variable	1	2	3	4	5
withinss	4.1	0.7	10.6	3.9	23.9
size	5	3	9	4	11
mpg*	-0.0582	-1.3701	-0.4391	-0.9110	1.0906
cyl*	-0.1050	1.0149	0.7660	1.0149	-1.2249
disp*	-0.5703	1.8284	0.5522	0.8857	-1.0133
hp*	-0.2696	1.0207	0.1374	1.8313	-0.9342
drat*	0.4777	-1.0160	-1.0763	0.1467	0.8872
wt*	-0.1924	2.1691	0.4020	0.3273	-0.9520
qsec*	-0.3429	-0.0609	0.0162	-1.5452	0.7211
(*) Centers
Total sum of squares: 217.0
Total within-cluster sum of squares: 43.2
Total between-cluster sum of squares: 173.8
BSS/TSS ratio: 80.1%
Number of iterations: 3

Separate column names into multiple rows

You can now separate collapsed column names into multiple rows with function separate_header(). This makes reporting aggregations calculated with ‘dplyr’ or ‘data.table’ easy.

In the following example we aggregate the famous ‘penguins’ table.

species	island	bill_length_avg	bill_length_sd	bill_depth_avg	bill_depth_sd	flipper_length_avg	flipper_length_sd
Adelie	Biscoe	39.0	2.5	18.4	1.2	188.8	6.7
Adelie	Dream	38.5	2.5	18.3	1.1	189.7	6.6
Adelie	Torgersen	39.0	3.0	18.4	1.3	191.2	6.2
Chinstrap	Dream	48.8	3.3	18.4	1.1	195.8	7.1
Gentoo	Biscoe	47.5	3.1	15.0	1.0	217.2	6.5

Without any intermediate manipulation, we create the ‘flextable’ and restructure header labels by using separate_header().

ft <- dat |> 
  flextable() |> 
  separate_header() |> 
  autofit()
ft

species	island	bill	flipper
Adelie	Biscoe	38.97500	2.480916	18.37045	1.1888199	188.7955	6.729247
Adelie	Dream	38.50179	2.465359	18.25179	1.1336171	189.7321	6.585083
Adelie	Torgersen	38.95098	3.025318	18.42941	1.3394468	191.1961	6.232238
Chinstrap	Dream	48.83382	3.339256	18.42059	1.1353951	195.8235	7.131894
Gentoo	Biscoe	47.50488	3.081857	14.98211	0.9812198	217.1870	6.484976

It’s then quite easy to customize it and produce a flextable as below:

species	island	bill	flipper
Adelie	Biscoe	38.98	2.48	18.37	1.19	188.80	6.73
Dream	38.50	2.47	18.25	1.13	189.73	6.59
Torgersen	38.95	3.03	18.43	1.34	191.20	6.23
Chinstrap	Dream	48.83	3.34	18.42	1.14	195.82	7.13
Gentoo	Biscoe	47.50	3.08	14.98	0.98	217.19	6.48
1Arithmetic Mean
2Standard Deviation

The code for this example is available in the flextable gallery here.

GitHub Discussions

Let us know about your issues, questions and feedbacks in ‘GitHub Discussions’ now available in the Github page of ‘flextable’: https://github.com/davidgohel/flextable/discussions.

variable	stat	nonchilled		chilled		Overall
variable	stat	Est.	N	Est.	N	Est.	N
Type	Quebec		21 (50.0 %)		21 (50.0 %)		42 (50.0 %)
	Mississippi		21 (50.0 %)		21 (50.0 %)		42 (50.0 %)
	Missing		0 (0.0 %)		0 (0.0 %)		0 (0.0 %)
conc	Min.	95.0		95.0		95.0
	1st Qu.	175.0		175.0		175.0
	Mean	435.0		435.0		435.0
	Median	350.0		350.0		350.0
	3rd Qu.	675.0		675.0		675.0
	Max.	1 000.0		1 000.0		1 000.0
	Missing		0 (0.0 %)		0 (0.0 %)		0 (0.0 %)
uptake	Min.	10.6		7.7		7.7
	1st Qu.	26.5		14.5		17.9
	Mean	30.6		23.8		27.2
	Median	31.3		19.7		28.3
	3rd Qu.	38.7		34.9		37.1
	Max.	45.5		42.4		45.5
	Missing		0 (0.0 %)		0 (0.0 %)		0 (0.0 %)

Subset of original dataset
cut	color	I1 (n = 180)	SI1 (n = 4800)	VS2 (n = 3569)
Good n = 2634	D	3.5 (± 2.2)	3.0 (± 2.8)	3.6 (± 3.4)
	E	4.4 (± 2.5)	3.2 (± 3.1)	3.8 (± 3.3)
	F	2.6 (± 1.9)	3.3 (± 3.0)	3.8 (± 3.2)
	G	3.2 (± 2.0)	4.1 (± 3.9)	4.1 (± 3.3)
	H	3.8 (± 2.2)	4.2 (± 3.8)	4.4 (± 4.2)
	I	4.2 (± 2.9)	4.7 (± 4.5)	6.0 (± 5.1)
	J	3.8 (± 2.1)	4.6 (± 3.9)	4.8 (± 3.6)
Very Good n = 5915	D	2.6 (± 0.8)	3.2 (± 2.9)	3.1 (± 3.3)
	E	3.4 (± 2.2)	3.2 (± 3.1)	3.3 (± 3.5)
	F	4.3 (± 2.6)	3.6 (± 3.2)	4.0 (± 3.8)
	G	3.2 (± 2.1)	3.5 (± 3.6)	4.4 (± 4.0)
	H	5.3 (± 2.5)	4.9 (± 4.4)	4.6 (± 4.0)
	I	6.0 (± 4.9)	5.2 (± 4.5)	5.8 (± 5.1)
	J	4.5 (± 2.7)	5.0 (± 4.1)	5.3 (± 4.5)